#1268 fix: update redirected_url to current page URL and enhance normalize_url function

Fix(auth): Fixed Docker JWT authentication

CHANGELOG.md

@@ -5,6 +5,16 @@ All notable changes to Crawl4AI will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [Unreleased]
+
+### Added
+- **🔒 HTTPS Preservation for Internal Links**: New `preserve_https_for_internal_links` configuration flag
+  - Maintains HTTPS scheme for internal links even when servers redirect to HTTP
+  - Prevents security downgrades during deep crawling
+  - Useful for security-conscious crawling and sites supporting both protocols
+  - Fully backward compatible with opt-in flag (default: `False`)
+  - Fixes issue #1410 where HTTPS URLs were being downgraded to HTTP
+
 ## [0.7.3] - 2025-08-09
 
 ### Added

crawl4ai/async_configs.py

@@ -97,13 +97,16 @@ def to_serializable_dict(obj: Any, ignore_default_value : bool = False) -> Dict:
                 if value != param.default and not ignore_default_value:
                     current_values[name] = to_serializable_dict(value)
         
-        if hasattr(obj, '__slots__'):
-            for slot in obj.__slots__:
-                if slot.startswith('_'):  # Handle private slots
-                    attr_name = slot[1:]  # Remove leading '_'
-                    value = getattr(obj, slot, None)
-                    if value is not None:
-                        current_values[attr_name] = to_serializable_dict(value)
+        # Don't serialize private __slots__ - they're internal implementation details
+        # not constructor parameters. This was causing URLPatternFilter to fail
+        # because _simple_suffixes was being serialized as 'simple_suffixes'
+        # if hasattr(obj, '__slots__'):
+        #     for slot in obj.__slots__:
+        #         if slot.startswith('_'):  # Handle private slots
+        #             attr_name = slot[1:]  # Remove leading '_'
+        #             value = getattr(obj, slot, None)
+        #             if value is not None:
+        #                 current_values[attr_name] = to_serializable_dict(value)
 
             
         
@@ -831,12 +834,6 @@ class HTTPCrawlerConfig:
         return HTTPCrawlerConfig.from_kwargs(config)
 
 class CrawlerRunConfig():
-    _UNWANTED_PROPS = {
-        'disable_cache' : 'Instead, use cache_mode=CacheMode.DISABLED',
-        'bypass_cache' : 'Instead, use cache_mode=CacheMode.BYPASS',
-        'no_cache_read' : 'Instead, use cache_mode=CacheMode.WRITE_ONLY',
-        'no_cache_write' : 'Instead, use cache_mode=CacheMode.READ_ONLY',
-    }
 
     """
     Configuration class for controlling how the crawler runs each crawl operation.
@@ -1043,6 +1040,12 @@ class CrawlerRunConfig():
 
         url: str = None  # This is not a compulsory parameter
     """
+    _UNWANTED_PROPS = {
+        'disable_cache' : 'Instead, use cache_mode=CacheMode.DISABLED',
+        'bypass_cache' : 'Instead, use cache_mode=CacheMode.BYPASS',
+        'no_cache_read' : 'Instead, use cache_mode=CacheMode.WRITE_ONLY',
+        'no_cache_write' : 'Instead, use cache_mode=CacheMode.READ_ONLY',
+    }
 
     def __init__(
         self,
@@ -1121,6 +1124,7 @@ class CrawlerRunConfig():
         exclude_domains: list = None,
         exclude_internal_links: bool = False,
         score_links: bool = False,
+        preserve_https_for_internal_links: bool = False,
         # Debugging and Logging Parameters
         verbose: bool = True,
         log_console: bool = False,
@@ -1244,6 +1248,7 @@ class CrawlerRunConfig():
         self.exclude_domains = exclude_domains or []
         self.exclude_internal_links = exclude_internal_links
         self.score_links = score_links
+        self.preserve_https_for_internal_links = preserve_https_for_internal_links
 
         # Debugging and Logging Parameters
         self.verbose = verbose
@@ -1517,6 +1522,7 @@ class CrawlerRunConfig():
             exclude_domains=kwargs.get("exclude_domains", []),
             exclude_internal_links=kwargs.get("exclude_internal_links", False),
             score_links=kwargs.get("score_links", False),
+            preserve_https_for_internal_links=kwargs.get("preserve_https_for_internal_links", False),
             # Debugging and Logging Parameters
             verbose=kwargs.get("verbose", True),
             log_console=kwargs.get("log_console", False),
@@ -1623,6 +1629,7 @@ class CrawlerRunConfig():
             "exclude_domains": self.exclude_domains,
             "exclude_internal_links": self.exclude_internal_links,
             "score_links": self.score_links,
+            "preserve_https_for_internal_links": self.preserve_https_for_internal_links,
             "verbose": self.verbose,
             "log_console": self.log_console,
             "capture_network_requests": self.capture_network_requests,

crawl4ai/async_crawler_strategy.py

@@ -1037,7 +1037,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                 downloaded_files=(
                     self._downloaded_files if self._downloaded_files else None
                 ),
-                redirected_url=redirected_url,
+                redirected_url=page.url,  # Update to current URL in case of JavaScript navigation
                 # Include captured data if enabled
                 network_requests=captured_requests if config.capture_network_requests else None,
                 console_messages=captured_console if config.capture_console_messages else None,

crawl4ai/async_webcrawler.py

@@ -354,6 +354,7 @@ class AsyncWebCrawler:
                     ###############################################################
                     # Process the HTML content, Call CrawlerStrategy.process_html #
                     ###############################################################
+                    from urllib.parse import urlparse
                     crawl_result: CrawlResult = await self.aprocess_html(
                         url=url,
                         html=html,
@@ -364,6 +365,7 @@ class AsyncWebCrawler:
                         verbose=config.verbose,
                         is_raw_html=True if url.startswith("raw:") else False,
                         redirected_url=async_response.redirected_url,
+                        original_scheme=urlparse(url).scheme,
                         **kwargs,
                     )
 
@@ -478,7 +480,7 @@ class AsyncWebCrawler:
             # Scraping Strategy Execution  #
             ################################
             result: ScrapingResult = scraping_strategy.scrap(
-                url, html, **params)
+                kwargs.get("redirected_url", url), html, **params)
 
             if result is None:
                 raise ValueError(

crawl4ai/content_scraping_strategy.py

@@ -258,7 +258,11 @@ class LXMLWebScrapingStrategy(ContentScrapingStrategy):
                 continue
 
             try:
-                normalized_href = normalize_url(href, url)
+                normalized_href = normalize_url(
+                    href, url,
+                    preserve_https=kwargs.get('preserve_https_for_internal_links', False),
+                    original_scheme=kwargs.get('original_scheme')
+                )
                 link_data = {
                     "href": normalized_href,
                     "text": link.text_content().strip(),

crawl4ai/deep_crawling/bff_strategy.py

@@ -47,7 +47,13 @@ class BestFirstCrawlingStrategy(DeepCrawlStrategy):
         self.url_scorer = url_scorer
         self.include_external = include_external
         self.max_pages = max_pages
-        self.logger = logger or logging.getLogger(__name__)
+        # self.logger = logger or logging.getLogger(__name__)
+        # Ensure logger is always a Logger instance, not a dict from serialization
+        if isinstance(logger, logging.Logger):
+            self.logger = logger
+        else:
+            # Create a new logger if logger is None, dict, or any other non-Logger type
+            self.logger = logging.getLogger(__name__)
         self.stats = TraversalStats(start_time=datetime.now())
         self._cancel_event = asyncio.Event()
         self._pages_crawled = 0

crawl4ai/deep_crawling/bfs_strategy.py

@@ -38,7 +38,13 @@ class BFSDeepCrawlStrategy(DeepCrawlStrategy):
         self.include_external = include_external
         self.score_threshold = score_threshold
         self.max_pages = max_pages
-        self.logger = logger or logging.getLogger(__name__)
+        # self.logger = logger or logging.getLogger(__name__)
+        # Ensure logger is always a Logger instance, not a dict from serialization
+        if isinstance(logger, logging.Logger):
+            self.logger = logger
+        else:
+            # Create a new logger if logger is None, dict, or any other non-Logger type
+            self.logger = logging.getLogger(__name__)
         self.stats = TraversalStats(start_time=datetime.now())
         self._cancel_event = asyncio.Event()
         self._pages_crawled = 0

crawl4ai/deep_crawling/filters.py

@@ -120,6 +120,9 @@ class URLPatternFilter(URLFilter):
     """Pattern filter balancing speed and completeness"""
 
     __slots__ = (
+        "patterns",  # Store original patterns for serialization
+        "use_glob",  # Store original use_glob for serialization  
+        "reverse",   # Store original reverse for serialization
         "_simple_suffixes",
         "_simple_prefixes",
         "_domain_patterns",
@@ -142,6 +145,11 @@ class URLPatternFilter(URLFilter):
         reverse: bool = False,
     ):
         super().__init__()
+        # Store original constructor params for serialization
+        self.patterns = patterns
+        self.use_glob = use_glob
+        self.reverse = reverse
+        
         self._reverse = reverse
         patterns = [patterns] if isinstance(patterns, (str, Pattern)) else patterns
 

crawl4ai/models.py

@@ -253,6 +253,16 @@ class CrawlResult(BaseModel):
         requirements change, this is where you would update the logic.
         """
         result = super().model_dump(*args, **kwargs)
+        
+        # Remove any property descriptors that might have been included
+        # These deprecated properties should not be in the serialized output
+        for key in ['fit_html', 'fit_markdown', 'markdown_v2']:
+            if key in result and isinstance(result[key], property):
+                # del result[key]
+                # Nasrin: I decided to convert it to string instead of removing it.
+                result[key] = str(result[key])
+        
+        # Add the markdown field properly
         if self._markdown is not None:
             result["markdown"] = self._markdown.model_dump() 
         return result

crawl4ai/utils.py

@@ -1790,6 +1790,10 @@ def perform_completion_with_backoff(
         except RateLimitError as e:
             print("Rate limit error:", str(e))
 
+            if attempt == max_attempts - 1:
+                # Last attempt failed, raise the error.
+                raise
+
             # Check if we have exhausted our max attempts
             if attempt < max_attempts - 1:
                 # Calculate the delay and wait
@@ -2145,8 +2149,12 @@ def normalize_url(
     *,
     drop_query_tracking=True,
     sort_query=True,
-    keep_fragment=False,
-    extra_drop_params=None
+    keep_fragment=True,
+    remove_fragments=None,  # alias for keep_fragment=False
+    extra_drop_params=None,
+    params_to_remove=None,  # alias for extra_drop_params
+    preserve_https=False,
+    original_scheme=None
 ):
     """
     Extended URL normalizer
@@ -2169,19 +2177,46 @@ def normalize_url(
     Returns
     -------
     str | None
-        A clean, canonical URL or None if href is empty/None.
+        A clean, canonical URL or the base URL if href is empty/None.
     """
     if not href:
-        return None
+        # For empty href, return the base URL (matching urljoin behavior)
+        return base_url
+
+    # Validate base URL format
+    parsed_base = urlparse(base_url)
+    if not parsed_base.scheme or not parsed_base.netloc:
+        raise ValueError(f"Invalid base URL format: {base_url}")
+    
+    if parsed_base.scheme.lower() not in ["http", "https"]:
+        # Handle special protocols
+        raise ValueError(f"Invalid base URL format: {base_url}")
 
     # Resolve relative paths first
     full_url = urljoin(base_url, href.strip())
+    
+    # Preserve HTTPS if requested and original scheme was HTTPS
+    if preserve_https and original_scheme == 'https':
+        parsed_full = urlparse(full_url)
+        parsed_base = urlparse(base_url)
+        # Only preserve HTTPS for same-domain links (not protocol-relative URLs)
+        # Protocol-relative URLs (//example.com) should follow the base URL's scheme
+        if (parsed_full.scheme == 'http' and 
+            parsed_full.netloc == parsed_base.netloc and
+            not href.strip().startswith('//')):
+            full_url = full_url.replace('http://', 'https://', 1)
 
     # Parse once, edit parts, then rebuild
     parsed = urlparse(full_url)
 
     # ── netloc ──
     netloc = parsed.netloc.lower()
+    
+    # Remove default ports (80 for http, 443 for https)
+    if ':' in netloc:
+        host, port = netloc.rsplit(':', 1)
+        if (parsed.scheme == 'http' and port == '80') or (parsed.scheme == 'https' and port == '443'):
+            netloc = host
 
     # ── path ──
     # Strip duplicate slashes and trailing "/" (except root)
@@ -2189,7 +2224,17 @@ def normalize_url(
     # The path from urlparse is already properly encoded
     path = parsed.path
     if path.endswith('/') and path != '/':
-        path = path.rstrip('/')
+        # Only strip trailing slash if the original href didn't have a trailing slash
+        # and the base_url didn't end with a slash
+        base_parsed = urlparse(base_url)
+        if not href.strip().endswith('/') and not base_parsed.path.endswith('/'):
+            path = path.rstrip('/')
+    # Add trailing slash for URLs without explicit paths (indicates directory)
+    # But skip this for special protocols that don't use standard URL structure
+    elif not path:
+        special_protocols = {"javascript:", "mailto:", "tel:", "file:", "data:"}
+        if not any(href.strip().lower().startswith(p) for p in special_protocols):
+            path = '/'
 
     # ── query ──
     query = parsed.query
@@ -2204,6 +2249,8 @@ def normalize_url(
             }
             if extra_drop_params:
                 default_tracking |= {p.lower() for p in extra_drop_params}
+            if params_to_remove:
+                default_tracking |= {p.lower() for p in params_to_remove}
             params = [(k, v) for k, v in params if k not in default_tracking]
 
         if sort_query:
@@ -2212,7 +2259,10 @@ def normalize_url(
         query = urlencode(params, doseq=True) if params else ''
 
     # ── fragment ──
-    fragment = parsed.fragment if keep_fragment else ''
+    if remove_fragments is True:
+        fragment = ''
+    else:
+        fragment = parsed.fragment if keep_fragment else ''
 
     # Re-assemble
     normalized = urlunparse((
@@ -2227,7 +2277,7 @@ def normalize_url(
     return normalized
 
 
-def normalize_url_for_deep_crawl(href, base_url):
+def normalize_url_for_deep_crawl(href, base_url, preserve_https=False, original_scheme=None):
     """Normalize URLs to ensure consistent format"""
     from urllib.parse import urljoin, urlparse, urlunparse, parse_qs, urlencode
 
@@ -2238,6 +2288,17 @@ def normalize_url_for_deep_crawl(href, base_url):
     # Use urljoin to handle relative URLs
     full_url = urljoin(base_url, href.strip())
     
+    # Preserve HTTPS if requested and original scheme was HTTPS
+    if preserve_https and original_scheme == 'https':
+        parsed_full = urlparse(full_url)
+        parsed_base = urlparse(base_url)
+        # Only preserve HTTPS for same-domain links (not protocol-relative URLs)
+        # Protocol-relative URLs (//example.com) should follow the base URL's scheme
+        if (parsed_full.scheme == 'http' and 
+            parsed_full.netloc == parsed_base.netloc and
+            not href.strip().startswith('//')):
+            full_url = full_url.replace('http://', 'https://', 1)
+    
     # Parse the URL for normalization
     parsed = urlparse(full_url)
     
@@ -2275,7 +2336,7 @@ def normalize_url_for_deep_crawl(href, base_url):
     return normalized
 
 @lru_cache(maxsize=10000)
-def efficient_normalize_url_for_deep_crawl(href, base_url):
+def efficient_normalize_url_for_deep_crawl(href, base_url, preserve_https=False, original_scheme=None):
     """Efficient URL normalization with proper parsing"""
     from urllib.parse import urljoin
     
@@ -2285,6 +2346,17 @@ def efficient_normalize_url_for_deep_crawl(href, base_url):
     # Resolve relative URLs
     full_url = urljoin(base_url, href.strip())
     
+    # Preserve HTTPS if requested and original scheme was HTTPS
+    if preserve_https and original_scheme == 'https':
+        parsed_full = urlparse(full_url)
+        parsed_base = urlparse(base_url)
+        # Only preserve HTTPS for same-domain links (not protocol-relative URLs)
+        # Protocol-relative URLs (//example.com) should follow the base URL's scheme
+        if (parsed_full.scheme == 'http' and 
+            parsed_full.netloc == parsed_base.netloc and
+            not href.strip().startswith('//')):
+            full_url = full_url.replace('http://', 'https://', 1)
+    
     # Use proper URL parsing
     parsed = urlparse(full_url)
     
@@ -2414,9 +2486,19 @@ def is_external_url(url: str, base_domain: str) -> bool:
         if not parsed.netloc:  # Relative URL
             return False
 
-        # Strip 'www.' from both domains for comparison
-        url_domain = parsed.netloc.lower().replace("www.", "")
-        base = base_domain.lower().replace("www.", "")
+        # Don't strip 'www.' from domains for comparison - treat www.example.com and example.com as different
+        url_domain = parsed.netloc.lower()
+        base = base_domain.lower()
+        
+        # Strip user credentials from URL domain
+        if '@' in url_domain:
+            url_domain = url_domain.split('@', 1)[1]
+        
+        # Strip ports from both for comparison (any port should be considered same domain)
+        if ':' in url_domain:
+            url_domain = url_domain.rsplit(':', 1)[0]
+        if ':' in base:
+            base = base.rsplit(':', 1)[0]
 
         # Check if URL domain ends with base domain
         return not url_domain.endswith(base)

deploy/docker/api.py

@@ -413,6 +413,9 @@ async def stream_results(crawler: AsyncWebCrawler, results_gen: AsyncGenerator)
                 server_memory_mb = _get_memory_mb()
                 result_dict = result.model_dump()
                 result_dict['server_memory_mb'] = server_memory_mb
+                # Ensure fit_html is JSON-serializable
+                if "fit_html" in result_dict and not (result_dict["fit_html"] is None or isinstance(result_dict["fit_html"], str)):
+                    result_dict["fit_html"] = None
                 # If PDF exists, encode it to base64
                 if result_dict.get('pdf') is not None:
                     result_dict['pdf'] = b64encode(result_dict['pdf']).decode('utf-8')
@@ -493,6 +496,9 @@ async def handle_crawl_request(
         processed_results = []
         for result in results:
             result_dict = result.model_dump()
+            # if fit_html is not a string, set it to None to avoid serialization errors
+            if "fit_html" in result_dict and not (result_dict["fit_html"] is None or isinstance(result_dict["fit_html"], str)):
+                result_dict["fit_html"] = None
             # If PDF exists, encode it to base64
             if result_dict.get('pdf') is not None:
                 result_dict['pdf'] = b64encode(result_dict['pdf']).decode('utf-8')

deploy/docker/auth.py

@@ -28,25 +28,43 @@ def create_access_token(data: dict, expires_delta: Optional[timedelta] = None) -
     signing_key = get_jwk_from_secret(SECRET_KEY)
     return instance.encode(to_encode, signing_key, alg='HS256')
 
-def verify_token(credentials: HTTPAuthorizationCredentials = Depends(security)) -> Dict:
+def verify_token(credentials: HTTPAuthorizationCredentials) -> Dict:
     """Verify the JWT token from the Authorization header."""
-
-    if credentials is None:
-        return None
+    
+    if not credentials or not credentials.credentials:
+        raise HTTPException(
+            status_code=401, 
+            detail="No token provided",
+            headers={"WWW-Authenticate": "Bearer"}
+        )
+    
     token = credentials.credentials
     verifying_key = get_jwk_from_secret(SECRET_KEY)
     try:
         payload = instance.decode(token, verifying_key, do_time_check=True, algorithms='HS256')
         return payload
-    except Exception:
-        raise HTTPException(status_code=401, detail="Invalid or expired token")
+    except Exception as e:
+        raise HTTPException(
+            status_code=401, 
+            detail=f"Invalid or expired token: {str(e)}",
+            headers={"WWW-Authenticate": "Bearer"}
+        )
 
 
 def get_token_dependency(config: Dict):
     """Return the token dependency if JWT is enabled, else a function that returns None."""
-
+    
     if config.get("security", {}).get("jwt_enabled", False):
-        return verify_token
+        def jwt_required(credentials: HTTPAuthorizationCredentials = Depends(security)) -> Dict:
+            """Enforce JWT authentication when enabled."""
+            if credentials is None:
+                raise HTTPException(
+                    status_code=401, 
+                    detail="Authentication required. Please provide a valid Bearer token.",
+                    headers={"WWW-Authenticate": "Bearer"}
+                )
+            return verify_token(credentials)
+        return jwt_required
     else:
         return lambda: None
 

deploy/docker/c4ai-doc-context.md

@@ -2241,7 +2241,7 @@ docker build -t crawl4ai
 
 | Argument | Description | Default | Options |
 |----------|-------------|---------|----------|
-| PYTHON_VERSION | Python version | 3.10 | 3.8, 3.9, 3.10 |
+| PYTHON_VERSION | Python version | 3.10 | 3.10, 3.11, 3.12, 3.13 |
 | INSTALL_TYPE | Feature set | default | default, all, torch, transformer |
 | ENABLE_GPU | GPU support | false | true, false |
 | APP_HOME | Install path | /app | any valid path |

deploy/docker/config.yml

@@ -38,8 +38,8 @@ rate_limiting:
 
 # Security Configuration
 security:
-  enabled: false 
-  jwt_enabled: false 
+  enabled: false
+  jwt_enabled: false
   https_redirect: false
   trusted_hosts: ["*"]
   headers:

deploy/docker/server.py

@@ -267,12 +267,26 @@ async def generate_html(
     Use when you need sanitized HTML structures for building schemas or further processing.
     """
     cfg = CrawlerRunConfig()
-    async with AsyncWebCrawler(config=BrowserConfig()) as crawler:
-        results = await crawler.arun(url=body.url, config=cfg)
-    raw_html = results[0].html
-    from crawl4ai.utils import preprocess_html_for_schema
-    processed_html = preprocess_html_for_schema(raw_html)
-    return JSONResponse({"html": processed_html, "url": body.url, "success": True})
+    try:
+        async with AsyncWebCrawler(config=BrowserConfig()) as crawler:
+            results = await crawler.arun(url=body.url, config=cfg)
+        # Check if the crawl was successful
+        if not results[0].success:
+            raise HTTPException(
+                status_code=500,
+                detail=results[0].error_message or "Crawl failed"
+            )
+        
+        raw_html = results[0].html
+        from crawl4ai.utils import preprocess_html_for_schema
+        processed_html = preprocess_html_for_schema(raw_html)
+        return JSONResponse({"html": processed_html, "url": body.url, "success": True})
+    except Exception as e:
+        # Log and raise as HTTP 500 for other exceptions
+        raise HTTPException(
+            status_code=500,
+            detail=str(e)
+        )
 
 # Screenshot endpoint
 
@@ -290,18 +304,29 @@ async def generate_screenshot(
     Use when you need an image snapshot of the rendered page. Its recommened to provide an output path to save the screenshot.
     Then in result instead of the screenshot you will get a path to the saved file.
     """
-    cfg = CrawlerRunConfig(
-        screenshot=True, screenshot_wait_for=body.screenshot_wait_for)
-    async with AsyncWebCrawler(config=BrowserConfig()) as crawler:
-        results = await crawler.arun(url=body.url, config=cfg)
-    screenshot_data = results[0].screenshot
-    if body.output_path:
-        abs_path = os.path.abspath(body.output_path)
-        os.makedirs(os.path.dirname(abs_path), exist_ok=True)
-        with open(abs_path, "wb") as f:
-            f.write(base64.b64decode(screenshot_data))
-        return {"success": True, "path": abs_path}
-    return {"success": True, "screenshot": screenshot_data}
+    try:
+        cfg = CrawlerRunConfig(
+            screenshot=True, screenshot_wait_for=body.screenshot_wait_for)
+        async with AsyncWebCrawler(config=BrowserConfig()) as crawler:
+            results = await crawler.arun(url=body.url, config=cfg)
+        if not results[0].success:
+            raise HTTPException(
+                status_code=500,
+                detail=results[0].error_message or "Crawl failed"
+            )
+        screenshot_data = results[0].screenshot
+        if body.output_path:
+            abs_path = os.path.abspath(body.output_path)
+            os.makedirs(os.path.dirname(abs_path), exist_ok=True)
+            with open(abs_path, "wb") as f:
+                f.write(base64.b64decode(screenshot_data))
+            return {"success": True, "path": abs_path}
+        return {"success": True, "screenshot": screenshot_data}
+    except Exception as e:
+        raise HTTPException(
+            status_code=500,
+            detail=str(e)
+        )
 
 # PDF endpoint
 
@@ -319,17 +344,28 @@ async def generate_pdf(
     Use when you need a printable or archivable snapshot of the page. It is recommended to provide an output path to save the PDF.
     Then in result instead of the PDF you will get a path to the saved file.
     """
-    cfg = CrawlerRunConfig(pdf=True)
-    async with AsyncWebCrawler(config=BrowserConfig()) as crawler:
-        results = await crawler.arun(url=body.url, config=cfg)
-    pdf_data = results[0].pdf
-    if body.output_path:
-        abs_path = os.path.abspath(body.output_path)
-        os.makedirs(os.path.dirname(abs_path), exist_ok=True)
-        with open(abs_path, "wb") as f:
-            f.write(pdf_data)
-        return {"success": True, "path": abs_path}
-    return {"success": True, "pdf": base64.b64encode(pdf_data).decode()}
+    try:
+        cfg = CrawlerRunConfig(pdf=True)
+        async with AsyncWebCrawler(config=BrowserConfig()) as crawler:
+            results = await crawler.arun(url=body.url, config=cfg)
+        if not results[0].success:
+            raise HTTPException(
+                status_code=500,
+                detail=results[0].error_message or "Crawl failed"
+            )
+        pdf_data = results[0].pdf
+        if body.output_path:
+            abs_path = os.path.abspath(body.output_path)
+            os.makedirs(os.path.dirname(abs_path), exist_ok=True)
+            with open(abs_path, "wb") as f:
+                f.write(pdf_data)
+            return {"success": True, "path": abs_path}
+        return {"success": True, "pdf": base64.b64encode(pdf_data).decode()}
+    except Exception as e:
+        raise HTTPException(
+            status_code=500,
+            detail=str(e)
+        )
 
 
 @app.post("/execute_js")
@@ -385,12 +421,23 @@ async def execute_js(
         ```
 
     """
-    cfg = CrawlerRunConfig(js_code=body.scripts)
-    async with AsyncWebCrawler(config=BrowserConfig()) as crawler:
-        results = await crawler.arun(url=body.url, config=cfg)
-    # Return JSON-serializable dict of the first CrawlResult
-    data = results[0].model_dump()
-    return JSONResponse(data)
+    try:
+        cfg = CrawlerRunConfig(js_code=body.scripts)
+        async with AsyncWebCrawler(config=BrowserConfig()) as crawler:
+            results = await crawler.arun(url=body.url, config=cfg)
+        if not results[0].success:
+            raise HTTPException(
+                status_code=500,
+                detail=results[0].error_message or "Crawl failed"
+            )
+        # Return JSON-serializable dict of the first CrawlResult
+        data = results[0].model_dump()
+        return JSONResponse(data)
+    except Exception as e:
+        raise HTTPException(
+            status_code=500,
+            detail=str(e)
+        )
 
 
 @app.get("/llm/{url:path}")
@@ -438,13 +485,16 @@ async def crawl(
     """
     if not crawl_request.urls:
         raise HTTPException(400, "At least one URL required")
-    res = await handle_crawl_request(
+    results = await handle_crawl_request(
         urls=crawl_request.urls,
         browser_config=crawl_request.browser_config,
         crawler_config=crawl_request.crawler_config,
         config=config,
     )
-    return JSONResponse(res)
+    # check if all of the results are not successful
+    if all(not result["success"] for result in results["results"]):
+        raise HTTPException(500, f"Crawl request failed: {results['results'][0]['error_message']}")
+    return JSONResponse(results)
 
 
 @app.post("/crawl/stream")

docs/examples/website-to-api/.gitignore

@@ -0,0 +1,221 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+#poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+#pdm.lock
+#pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+#pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# Redis 
+*.rdb
+*.aof
+*.pid
+
+# RabbitMQ
+mnesia/
+rabbitmq/
+rabbitmq-data/
+
+# ActiveMQ
+activemq-data/
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# Abstra
+# Abstra is an AI-powered process automation framework.
+# Ignore directories containing user credentials, local state, and settings.
+# Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#  Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#  that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#  and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#  you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+
+# Streamlit
+.streamlit/secrets.toml
+
+#directories
+models
+schemas
+saved_requests

docs/examples/website-to-api/README.md

@@ -0,0 +1,252 @@
+# Web Scraper API with Custom Model Support
+
+A powerful web scraping API that converts any website into structured data using AI. Features a beautiful minimalist frontend interface and support for custom LLM models!
+
+## Features
+
+- **AI-Powered Scraping**: Provide a URL and plain English query to extract structured data
+- **Beautiful Frontend**: Modern minimalist black-and-white interface with smooth UX
+- **Custom Model Support**: Use any LLM provider (OpenAI, Gemini, Anthropic, etc.) with your own API keys
+- **Model Management**: Save, list, and manage multiple model configurations via web interface
+- **Dual Scraping Approaches**: Choose between Schema-based (faster) or LLM-based (more flexible) extraction
+- **API Request History**: Automatic saving and display of all API requests with cURL commands
+- **Schema Caching**: Intelligent caching of generated schemas for faster subsequent requests
+- **Duplicate Prevention**: Avoids saving duplicate requests (same URL + query)
+- **RESTful API**: Easy-to-use HTTP endpoints for all operations
+
+## Quick Start
+
+### 1. Install Dependencies
+
+```bash
+pip install -r requirements.txt
+```
+
+### 2. Start the API Server
+
+```bash
+python app.py
+```
+
+The server will start on `http://localhost:8000` with a beautiful web interface!
+
+### 3. Using the Web Interface
+
+Once the server is running, open your browser and go to `http://localhost:8000` to access the modern web interface!
+
+#### Pages:
+- **Scrape Data**: Enter URLs and queries to extract structured data
+- **Models**: Manage your AI model configurations (add, list, delete)
+- **API Requests**: View history of all scraping requests with cURL commands
+
+#### Features:
+- **Minimalist Design**: Clean black-and-white theme inspired by modern web apps
+- **Real-time Results**: See extracted data in formatted JSON
+- **Copy to Clipboard**: Easy copying of results
+- **Toast Notifications**: User-friendly feedback
+- **Dual Scraping Modes**: Choose between Schema-based and LLM-based approaches
+
+## Model Management
+
+### Adding Models via Web Interface
+
+1. Go to the **Models** page
+2. Enter your model details:
+   - **Provider**: LLM provider (e.g., `gemini/gemini-2.5-flash`, `openai/gpt-4o`)
+   - **API Token**: Your API key for the provider
+3. Click "Add Model"
+
+### API Usage for Model Management
+
+#### Save a Model Configuration
+
+```bash
+curl -X POST "http://localhost:8000/models" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "provider": "gemini/gemini-2.5-flash",
+    "api_token": "your-api-key-here"
+  }'
+```
+
+#### List Saved Models
+
+```bash
+curl -X GET "http://localhost:8000/models"
+```
+
+#### Delete a Model Configuration
+
+```bash
+curl -X DELETE "http://localhost:8000/models/my-gemini"
+```
+
+## Scraping Approaches
+
+### 1. Schema-based Scraping (Faster)
+- Generates CSS selectors for targeted extraction
+- Caches schemas for repeated requests
+- Faster execution for structured websites
+
+### 2. LLM-based Scraping (More Flexible)
+- Direct LLM extraction without schema generation
+- More flexible for complex or dynamic content
+- Better for unstructured data extraction
+
+## Supported LLM Providers
+
+The API supports any LLM provider that crawl4ai supports, including:
+
+- **Google Gemini**: `gemini/gemini-2.5-flash`, `gemini/gemini-pro`
+- **OpenAI**: `openai/gpt-4`, `openai/gpt-3.5-turbo`
+- **Anthropic**: `anthropic/claude-3-opus`, `anthropic/claude-3-sonnet`
+- **And more...**
+
+## API Endpoints
+
+### Core Endpoints
+
+- `POST /scrape` - Schema-based scraping
+- `POST /scrape-with-llm` - LLM-based scraping
+- `GET /schemas` - List cached schemas
+- `POST /clear-cache` - Clear schema cache
+- `GET /health` - Health check
+
+### Model Management Endpoints
+
+- `GET /models` - List saved model configurations
+- `POST /models` - Save a new model configuration
+- `DELETE /models/{model_name}` - Delete a model configuration
+
+### API Request History
+
+- `GET /saved-requests` - List all saved API requests
+- `DELETE /saved-requests/{request_id}` - Delete a saved request
+
+## Request/Response Examples
+
+### Scrape Request
+
+```json
+{
+  "url": "https://example.com",
+  "query": "Extract the product name, price, and description",
+  "model_name": "my-custom-model"
+}
+```
+
+### Scrape Response
+
+```json
+{
+  "success": true,
+  "url": "https://example.com",
+  "query": "Extract the product name, price, and description",
+  "extracted_data": {
+    "product_name": "Example Product",
+    "price": "$99.99",
+    "description": "This is an example product description"
+  },
+  "schema_used": { ... },
+  "timestamp": "2024-01-01T12:00:00Z"
+}
+```
+
+### Model Configuration Request
+
+```json
+{
+  "provider": "gemini/gemini-2.5-flash",
+  "api_token": "your-api-key-here"
+}
+```
+
+## Testing
+
+Run the test script to verify the model management functionality:
+
+```bash
+python test_models.py
+```
+
+## File Structure
+
+```
+parse_example/
+├── api_server.py          # FastAPI server with all endpoints
+├── web_scraper_lib.py     # Core scraping library
+├── test_models.py         # Test script for model management
+├── requirements.txt       # Dependencies
+├── static/               # Frontend files
+│   ├── index.html        # Main HTML interface
+│   ├── styles.css        # CSS styles (minimalist theme)
+│   └── script.js         # JavaScript functionality
+├── schemas/              # Cached schemas
+├── models/               # Saved model configurations
+├── saved_requests/       # API request history
+└── README.md            # This file
+```
+
+## Advanced Usage
+
+### Using the Library Directly
+
+```python
+from web_scraper_lib import WebScraperAgent
+
+# Initialize agent
+agent = WebScraperAgent()
+
+# Save a model configuration
+agent.save_model_config(
+    model_name="my-model",
+    provider="openai/gpt-4",
+    api_token="your-api-key"
+)
+
+# Schema-based scraping
+result = await agent.scrape_data(
+    url="https://example.com",
+    query="Extract product information",
+    model_name="my-model"
+)
+
+# LLM-based scraping
+result = await agent.scrape_data_with_llm(
+    url="https://example.com",
+    query="Extract product information",
+    model_name="my-model"
+)
+```
+
+### Schema Caching
+
+The system automatically caches generated schemas based on URL and query combinations:
+
+- **First request**: Generates schema using AI
+- **Subsequent requests**: Uses cached schema for faster extraction
+
+### API Request History
+
+All API requests are automatically saved with:
+- Request details (URL, query, model used)
+- Response data
+- Timestamp
+- cURL command for re-execution
+
+### Duplicate Prevention
+
+The system prevents saving duplicate requests:
+- Same URL + query combinations are not saved multiple times
+- Returns existing request ID for duplicates
+- Keeps the API request history clean
+
+## Error Handling
+
+The API provides detailed error messages for common issues:
+
+- Invalid URLs
+- Missing model configurations
+- API key errors
+- Network timeouts
+- Parsing errors

docs/examples/website-to-api/api_server.py

@@ -0,0 +1,363 @@
+from fastapi import FastAPI, HTTPException
+from fastapi.staticfiles import StaticFiles
+from fastapi.responses import FileResponse
+from pydantic import BaseModel, HttpUrl
+from typing import Dict, Any, Optional, Union, List
+import uvicorn
+import asyncio
+import os
+import json
+from datetime import datetime
+from web_scraper_lib import WebScraperAgent, scrape_website
+
+app = FastAPI(
+    title="Web Scraper API",
+    description="Convert any website into a structured data API. Provide a URL and tell AI what data you need in plain English.",
+    version="1.0.0"
+)
+
+# Mount static files
+if os.path.exists("static"):
+    app.mount("/static", StaticFiles(directory="static"), name="static")
+
+# Mount assets directory
+if os.path.exists("assets"):
+    app.mount("/assets", StaticFiles(directory="assets"), name="assets")
+
+# Initialize the scraper agent
+scraper_agent = WebScraperAgent()
+
+# Create directory for saved API requests
+os.makedirs("saved_requests", exist_ok=True)
+
+class ScrapeRequest(BaseModel):
+    url: HttpUrl
+    query: str
+    model_name: Optional[str] = None
+
+class ModelConfigRequest(BaseModel):
+    model_name: str
+    provider: str
+    api_token: str
+
+class ScrapeResponse(BaseModel):
+    success: bool
+    url: str
+    query: str
+    extracted_data: Union[Dict[str, Any], list]
+    schema_used: Optional[Dict[str, Any]] = None
+    timestamp: Optional[str] = None
+    error: Optional[str] = None
+
+class SavedApiRequest(BaseModel):
+    id: str
+    endpoint: str
+    method: str
+    headers: Dict[str, str]
+    body: Dict[str, Any]
+    timestamp: str
+    response: Optional[Dict[str, Any]] = None
+
+def save_api_request(endpoint: str, method: str, headers: Dict[str, str], body: Dict[str, Any], response: Optional[Dict[str, Any]] = None) -> str:
+    """Save an API request to a JSON file."""
+    
+    # Check for duplicate requests (same URL and query)
+    if endpoint in ["/scrape", "/scrape-with-llm"] and "url" in body and "query" in body:
+        existing_requests = get_saved_requests()
+        for existing_request in existing_requests:
+            if (existing_request.endpoint == endpoint and 
+                existing_request.body.get("url") == body["url"] and 
+                existing_request.body.get("query") == body["query"]):
+                print(f"Duplicate request found for URL: {body['url']} and query: {body['query']}")
+                return existing_request.id  # Return existing request ID instead of creating new one
+    
+    request_id = datetime.now().strftime("%Y%m%d_%H%M%S_%f")[:-3]
+    
+    saved_request = SavedApiRequest(
+        id=request_id,
+        endpoint=endpoint,
+        method=method,
+        headers=headers,
+        body=body,
+        timestamp=datetime.now().isoformat(),
+        response=response
+    )
+    
+    file_path = os.path.join("saved_requests", f"{request_id}.json")
+    with open(file_path, "w") as f:
+        json.dump(saved_request.dict(), f, indent=2)
+    
+    return request_id
+
+def get_saved_requests() -> List[SavedApiRequest]:
+    """Get all saved API requests."""
+    requests = []
+    if os.path.exists("saved_requests"):
+        for filename in os.listdir("saved_requests"):
+            if filename.endswith('.json'):
+                file_path = os.path.join("saved_requests", filename)
+                try:
+                    with open(file_path, "r") as f:
+                        data = json.load(f)
+                        requests.append(SavedApiRequest(**data))
+                except Exception as e:
+                    print(f"Error loading saved request {filename}: {e}")
+    
+    # Sort by timestamp (newest first)
+    requests.sort(key=lambda x: x.timestamp, reverse=True)
+    return requests
+
+@app.get("/")
+async def root():
+    """Serve the frontend interface."""
+    if os.path.exists("static/index.html"):
+        return FileResponse("static/index.html")
+    else:
+        return {
+            "message": "Web Scraper API",
+            "description": "Convert any website into structured data with AI",
+            "endpoints": {
+                "/scrape": "POST - Scrape data from a website",
+                "/schemas": "GET - List cached schemas",
+                "/clear-cache": "POST - Clear schema cache",
+                "/models": "GET - List saved model configurations",
+                "/models": "POST - Save a new model configuration",
+                "/models/{model_name}": "DELETE - Delete a model configuration",
+                "/saved-requests": "GET - List saved API requests"
+            }
+        }
+
+@app.post("/scrape", response_model=ScrapeResponse)
+async def scrape_website_endpoint(request: ScrapeRequest):
+    """
+    Scrape structured data from any website.
+    
+    This endpoint:
+    1. Takes a URL and plain English query
+    2. Generates a custom scraper using AI
+    3. Returns structured data
+    """
+    try:
+        # Save the API request
+        headers = {"Content-Type": "application/json"}
+        body = {
+            "url": str(request.url),
+            "query": request.query,
+            "model_name": request.model_name
+        }
+        
+        result = await scraper_agent.scrape_data(
+            url=str(request.url),
+            query=request.query,
+            model_name=request.model_name
+        )
+        
+        response_data = ScrapeResponse(
+            success=True,
+            url=result["url"],
+            query=result["query"],
+            extracted_data=result["extracted_data"],
+            schema_used=result["schema_used"],
+            timestamp=result["timestamp"]
+        )
+        
+        # Save the request with response
+        save_api_request(
+            endpoint="/scrape",
+            method="POST",
+            headers=headers,
+            body=body,
+            response=response_data.dict()
+        )
+        
+        return response_data
+    
+    except Exception as e:
+        # Save the failed request
+        headers = {"Content-Type": "application/json"}
+        body = {
+            "url": str(request.url),
+            "query": request.query,
+            "model_name": request.model_name
+        }
+        
+        save_api_request(
+            endpoint="/scrape",
+            method="POST",
+            headers=headers,
+            body=body,
+            response={"error": str(e)}
+        )
+        
+        raise HTTPException(status_code=500, detail=f"Scraping failed: {str(e)}")
+
+@app.post("/scrape-with-llm", response_model=ScrapeResponse)
+async def scrape_website_endpoint_with_llm(request: ScrapeRequest):
+    """
+    Scrape structured data from any website using a custom LLM model.
+    """
+    try:
+        # Save the API request
+        headers = {"Content-Type": "application/json"}
+        body = {
+            "url": str(request.url),
+            "query": request.query,
+            "model_name": request.model_name
+        }
+        
+        result = await scraper_agent.scrape_data_with_llm(
+            url=str(request.url),
+            query=request.query,
+            model_name=request.model_name
+        )
+        
+        response_data = ScrapeResponse(
+            success=True,
+            url=result["url"],
+            query=result["query"],
+            extracted_data=result["extracted_data"],
+            timestamp=result["timestamp"]
+        )
+        
+        # Save the request with response
+        save_api_request(
+            endpoint="/scrape-with-llm",
+            method="POST",
+            headers=headers,
+            body=body,
+            response=response_data.dict()
+        )
+        
+        return response_data
+    
+    except Exception as e:
+        # Save the failed request
+        headers = {"Content-Type": "application/json"}
+        body = {
+            "url": str(request.url),
+            "query": request.query,
+            "model_name": request.model_name
+        }
+        
+        save_api_request(
+            endpoint="/scrape-with-llm",
+            method="POST",
+            headers=headers,
+            body=body,
+            response={"error": str(e)}
+        )
+        
+        raise HTTPException(status_code=500, detail=f"Scraping failed: {str(e)}")
+
+@app.get("/saved-requests")
+async def list_saved_requests():
+    """List all saved API requests."""
+    try:
+        requests = get_saved_requests()
+        return {
+            "success": True,
+            "requests": [req.dict() for req in requests],
+            "count": len(requests)
+        }
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to list saved requests: {str(e)}")
+
+@app.delete("/saved-requests/{request_id}")
+async def delete_saved_request(request_id: str):
+    """Delete a saved API request."""
+    try:
+        file_path = os.path.join("saved_requests", f"{request_id}.json")
+        if os.path.exists(file_path):
+            os.remove(file_path)
+            return {
+                "success": True,
+                "message": f"Saved request '{request_id}' deleted successfully"
+            }
+        else:
+            raise HTTPException(status_code=404, detail=f"Saved request '{request_id}' not found")
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to delete saved request: {str(e)}")
+
+@app.get("/schemas")
+async def list_cached_schemas():
+    """List all cached schemas."""
+    try:
+        schemas = await scraper_agent.get_cached_schemas()
+        return {
+            "success": True,
+            "cached_schemas": schemas,
+            "count": len(schemas)
+        }
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to list schemas: {str(e)}")
+
+@app.post("/clear-cache")
+async def clear_schema_cache():
+    """Clear all cached schemas."""
+    try:
+        scraper_agent.clear_cache()
+        return {
+            "success": True,
+            "message": "Schema cache cleared successfully"
+        }
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to clear cache: {str(e)}")
+
+@app.get("/models")
+async def list_models():
+    """List all saved model configurations."""
+    try:
+        models = scraper_agent.list_saved_models()
+        return {
+            "success": True,
+            "models": models,
+            "count": len(models)
+        }
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to list models: {str(e)}")
+
+@app.post("/models")
+async def save_model_config(request: ModelConfigRequest):
+    """Save a new model configuration."""
+    try:
+        success = scraper_agent.save_model_config(
+            model_name=request.model_name,
+            provider=request.provider,
+            api_token=request.api_token
+        )
+        
+        if success:
+            return {
+                "success": True,
+                "message": f"Model configuration '{request.model_name}' saved successfully"
+            }
+        else:
+            raise HTTPException(status_code=500, detail="Failed to save model configuration")
+    
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to save model: {str(e)}")
+
+@app.delete("/models/{model_name}")
+async def delete_model_config(model_name: str):
+    """Delete a model configuration."""
+    try:
+        success = scraper_agent.delete_model_config(model_name)
+        
+        if success:
+            return {
+                "success": True,
+                "message": f"Model configuration '{model_name}' deleted successfully"
+            }
+        else:
+            raise HTTPException(status_code=404, detail=f"Model configuration '{model_name}' not found")
+    
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to delete model: {str(e)}")
+
+@app.get("/health")
+async def health_check():
+    """Health check endpoint."""
+    return {"status": "healthy", "service": "web-scraper-api"}
+
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=8000)

docs/examples/website-to-api/app.py

@@ -0,0 +1,49 @@
+#!/usr/bin/env python3
+"""
+Startup script for the Web Scraper API with frontend interface.
+"""
+
+import os
+import sys
+import uvicorn
+from pathlib import Path
+
+def main():
+    # Check if static directory exists
+    static_dir = Path("static")
+    if not static_dir.exists():
+        print("❌ Static directory not found!")
+        print("Please make sure the 'static' directory exists with the frontend files.")
+        sys.exit(1)
+    
+    # Check if required frontend files exist
+    required_files = ["index.html", "styles.css", "script.js"]
+    missing_files = []
+    
+    for file in required_files:
+        if not (static_dir / file).exists():
+            missing_files.append(file)
+    
+    if missing_files:
+        print(f"❌ Missing frontend files: {', '.join(missing_files)}")
+        print("Please make sure all frontend files are present in the static directory.")
+        sys.exit(1)
+    
+    print("🚀 Starting Web Scraper API with Frontend Interface")
+    print("=" * 50)
+    print("📁 Static files found and ready to serve")
+    print("🌐 Frontend will be available at: http://localhost:8000")
+    print("🔌 API endpoints available at: http://localhost:8000/docs")
+    print("=" * 50)
+    
+    # Start the server
+    uvicorn.run(
+        "api_server:app",
+        host="0.0.0.0",
+        port=8000,
+        reload=True,
+        log_level="info"
+    )
+
+if __name__ == "__main__":
+    main() 

docs/examples/website-to-api/requirements.txt

@@ -0,0 +1,5 @@
+crawl4ai
+fastapi
+uvicorn
+pydantic
+litellm

docs/examples/website-to-api/static/index.html

@@ -0,0 +1,201 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Web2API Example</title>
+    <link rel="stylesheet" href="/static/styles.css">
+    <link href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0/css/all.min.css" rel="stylesheet">
+</head>
+<body>
+    <!-- Header -->
+    <header class="header">
+        <div class="header-content">
+            <div class="logo">
+                <img src="/assets/crawl4ai_logo.jpg" alt="Crawl4AI Logo" class="logo-image">
+                <span>Web2API Example</span>
+            </div>
+            <nav class="nav-links">
+                <a href="#" class="nav-link active" data-page="scrape">Scrape</a>
+                <a href="#" class="nav-link" data-page="models">Models</a>
+                <a href="#" class="nav-link" data-page="requests">API Requests</a>
+            </nav>
+        </div>
+    </header>
+
+    <!-- Main Content -->
+    <main class="main-content">
+        <!-- Scrape Page -->
+        <div id="scrape-page" class="page active">
+            <div class="hero-section">
+                <h1 class="hero-title">Turn Any Website Into An API</h1>
+                <p class="hero-subtitle">This example shows how to turn any website into an API using Crawl4AI.</p>
+            </div>
+
+            <!-- Workflow Demonstration -->
+            <div class="workflow-demo">
+                <div class="workflow-step">
+                    <h3 class="step-title">1. Your Request</h3>
+                    <div class="request-box">
+                        <div class="input-group">
+                            <label>URL:</label>
+                            <input type="url" id="url" name="url" placeholder="https://example-bookstore.com/new-releases" required>
+                        </div>
+                        <div class="input-group">
+                            <label>QUERY:</label>
+                            <textarea id="query" name="query" placeholder="Extract all the book titles, their authors, and the biography of the author" required></textarea>
+                        </div>
+                        <div class="form-options">
+                            <div class="option-group">
+                                <label for="scraping-approach">Approach:</label>
+                                <select id="scraping-approach" name="scraping_approach">
+                                    <option value="llm">LLM-based (More Flexible)</option>
+                                    <option value="schema">Schema-based (Uses LLM once!)</option>
+                                </select>
+                            </div>
+                            <div class="option-group">
+                                <label for="model-select">Model:</label>
+                                <select id="model-select" name="model_name" required>
+                                    <option value="">Select a Model</option>
+                                </select>
+                            </div>
+                        </div>
+                        <button type="submit" id="extract-btn" class="extract-btn">
+                            <i class="fas fa-magic"></i>
+                            Extract Data
+                        </button>
+                    </div>
+                </div>
+
+                <div class="workflow-arrow">→</div>
+
+                <div class="workflow-step">
+                    <h3 class="step-title">2. Your Instant API & Data</h3>
+                    <div class="response-container">
+                        <div class="api-request-box">
+                            <label>API Request (cURL):</label>
+                            <pre id="curl-example">curl -X POST http://localhost:8000/scrape -H "Content-Type: application/json" -d '{"url": "...", "query": "..."}'
+
+# Or for LLM-based approach:
+curl -X POST http://localhost:8000/scrape-with-llm -H "Content-Type: application/json" -d '{"url": "...", "query": "..."}'</pre>
+                        </div>
+                        <div class="json-response-box">
+                            <label>JSON Response:</label>
+                            <pre id="json-output">{
+  "success": true,
+  "extracted_data": [
+    {
+      "title": "Example Book",
+      "author": "John Doe",
+      "description": "A great book..."
+    }
+  ]
+}</pre>
+                        </div>
+                    </div>
+                </div>
+            </div>
+
+            <!-- Results Section -->
+            <div id="results-section" class="results-section" style="display: none;">
+                <div class="results-header">
+                    <h2>Extracted Data</h2>
+                    <button id="copy-json" class="copy-btn">
+                        <i class="fas fa-copy"></i>
+                        Copy JSON
+                    </button>
+                </div>
+                <div class="results-content">
+                    <div class="result-info">
+                        <div class="info-item">
+                            <span class="label">URL:</span>
+                            <span id="result-url" class="value"></span>
+                        </div>
+                        <div class="info-item">
+                            <span class="label">Query:</span>
+                            <span id="result-query" class="value"></span>
+                        </div>
+                        <div class="info-item">
+                            <span class="label">Model Used:</span>
+                            <span id="result-model" class="value"></span>
+                        </div>
+                    </div>
+                    <div class="json-display">
+                        <pre id="actual-json-output"></pre>
+                    </div>
+                </div>
+            </div>
+
+            <!-- Loading State -->
+            <div id="loading" class="loading" style="display: none;">
+                <div class="spinner"></div>
+                <p>AI is analyzing the website and extracting data...</p>
+            </div>
+        </div>
+
+        <!-- Models Page -->
+        <div id="models-page" class="page">
+            <div class="models-header">
+                <h1>Model Configuration</h1>
+                <p>Configure and manage your AI model configurations</p>
+            </div>
+
+            <div class="models-container">
+                <!-- Add New Model Form -->
+                <div class="model-form-section">
+                    <h3>Add New Model</h3>
+                    <form id="model-form" class="model-form">
+                        <div class="form-row">
+                            <div class="input-group">
+                                <label for="model-name">Model Name:</label>
+                                <input type="text" id="model-name" name="model_name" placeholder="my-gemini" required>
+                            </div>
+                            <div class="input-group">
+                                <label for="provider">Provider:</label>
+                                <input type="text" id="provider" name="provider" placeholder="gemini/gemini-2.5-flash" required>
+                            </div>
+                        </div>
+                        
+                        <div class="input-group">
+                            <label for="api-token">API Token:</label>
+                            <input type="password" id="api-token" name="api_token" placeholder="Enter your API token" required>
+                        </div>
+
+                        <button type="submit" class="save-btn">
+                            <i class="fas fa-save"></i>
+                            Save Model
+                        </button>
+                    </form>
+                </div>
+
+                <!-- Saved Models List -->
+                <div class="saved-models-section">
+                    <h3>Saved Models</h3>
+                    <div id="models-list" class="models-list">
+                        <!-- Models will be loaded here -->
+                    </div>
+                </div>
+            </div>
+        </div>
+
+        <!-- API Requests Page -->
+        <div id="requests-page" class="page">
+            <div class="requests-header">
+                <h1>Saved API Requests</h1>
+                <p>View and manage your previous API requests</p>
+            </div>
+
+            <div class="requests-container">
+                <div class="requests-list" id="requests-list">
+                    <!-- Saved requests will be loaded here -->
+                </div>
+            </div>
+        </div>
+    </main>
+
+    <!-- Toast Notifications -->
+    <div id="toast-container" class="toast-container"></div>
+
+    <script src="/static/script.js"></script>
+</body>
+</html> 

docs/examples/website-to-api/static/script.js

@@ -0,0 +1,401 @@
+// API Configuration
+const API_BASE_URL = 'http://localhost:8000';
+
+// DOM Elements
+const navLinks = document.querySelectorAll('.nav-link');
+const pages = document.querySelectorAll('.page');
+const scrapeForm = document.getElementById('scrape-form');
+const modelForm = document.getElementById('model-form');
+const modelSelect = document.getElementById('model-select');
+const modelsList = document.getElementById('models-list');
+const resultsSection = document.getElementById('results-section');
+const loadingSection = document.getElementById('loading');
+const copyJsonBtn = document.getElementById('copy-json');
+
+// Navigation
+navLinks.forEach(link => {
+    link.addEventListener('click', (e) => {
+        e.preventDefault();
+        const targetPage = link.dataset.page;
+        
+        // Update active nav link
+        navLinks.forEach(l => l.classList.remove('active'));
+        link.classList.add('active');
+        
+        // Show target page
+        pages.forEach(page => page.classList.remove('active'));
+        document.getElementById(`${targetPage}-page`).classList.add('active');
+        
+        // Load data for the page
+        if (targetPage === 'models') {
+            loadModels();
+        } else if (targetPage === 'requests') {
+            loadSavedRequests();
+        }
+    });
+});
+
+// Scrape Form Handler
+document.getElementById('extract-btn').addEventListener('click', async (e) => {
+    e.preventDefault();
+    
+    // Scroll to results section immediately when button is clicked
+    document.getElementById('results-section').scrollIntoView({ 
+        behavior: 'smooth', 
+        block: 'start' 
+    });
+    
+    const url = document.getElementById('url').value;
+    const query = document.getElementById('query').value;
+    const headless = true; // Always use headless mode
+    const model_name = document.getElementById('model-select').value || null;
+    const scraping_approach = document.getElementById('scraping-approach').value;
+    
+    if (!url || !query) {
+        showToast('Please fill in both URL and query fields', 'error');
+        return;
+    }
+    
+    if (!model_name) {
+        showToast('Please select a model from the dropdown or add one from the Models page', 'error');
+        return;
+    }
+    
+    const data = {
+        url: url,
+        query: query,
+        headless: headless,
+        model_name: model_name
+    };
+    
+    // Show loading state
+    showLoading(true);
+    hideResults();
+    
+    try {
+        // Choose endpoint based on scraping approach
+        const endpoint = scraping_approach === 'llm' ? '/scrape-with-llm' : '/scrape';
+        
+        const response = await fetch(`${API_BASE_URL}${endpoint}`, {
+            method: 'POST',
+            headers: {
+                'Content-Type': 'application/json'
+            },
+            body: JSON.stringify(data)
+        });
+        
+        const result = await response.json();
+        
+        if (response.ok) {
+            displayResults(result);
+            showToast(`Data extracted successfully using ${scraping_approach === 'llm' ? 'LLM-based' : 'Schema-based'} approach!`, 'success');
+        } else {
+            throw new Error(result.detail || 'Failed to extract data');
+        }
+    } catch (error) {
+        console.error('Scraping error:', error);
+        showToast(`Error: ${error.message}`, 'error');
+    } finally {
+        showLoading(false);
+    }
+});
+
+// Model Form Handler
+modelForm.addEventListener('submit', async (e) => {
+    e.preventDefault();
+    
+    const formData = new FormData(modelForm);
+    const data = {
+        model_name: formData.get('model_name'),
+        provider: formData.get('provider'),
+        api_token: formData.get('api_token')
+    };
+    
+    try {
+        const response = await fetch(`${API_BASE_URL}/models`, {
+            method: 'POST',
+            headers: {
+                'Content-Type': 'application/json'
+            },
+            body: JSON.stringify(data)
+        });
+        
+        const result = await response.json();
+        
+        if (response.ok) {
+            showToast('Model saved successfully!', 'success');
+            modelForm.reset();
+            loadModels();
+            loadModelSelect();
+        } else {
+            throw new Error(result.detail || 'Failed to save model');
+        }
+    } catch (error) {
+        console.error('Model save error:', error);
+        showToast(`Error: ${error.message}`, 'error');
+    }
+});
+
+// Copy JSON Button
+copyJsonBtn.addEventListener('click', () => {
+    const actualJsonOutput = document.getElementById('actual-json-output');
+    const textToCopy = actualJsonOutput.textContent;
+    
+    navigator.clipboard.writeText(textToCopy).then(() => {
+        showToast('JSON copied to clipboard!', 'success');
+    }).catch(() => {
+        showToast('Failed to copy JSON', 'error');
+    });
+});
+
+// Load Models
+async function loadModels() {
+    try {
+        const response = await fetch(`${API_BASE_URL}/models`);
+        const result = await response.json();
+        
+        if (response.ok) {
+            displayModels(result.models);
+        } else {
+            throw new Error(result.detail || 'Failed to load models');
+        }
+    } catch (error) {
+        console.error('Load models error:', error);
+        showToast(`Error: ${error.message}`, 'error');
+    }
+}
+
+// Display Models
+function displayModels(models) {
+    if (models.length === 0) {
+        modelsList.innerHTML = '<p style="text-align: center; color: #7f8c8d; padding: 2rem;">No models saved yet. Add your first model above!</p>';
+        return;
+    }
+    
+    modelsList.innerHTML = models.map(model => `
+        <div class="model-card">
+            <div class="model-info">
+                <div class="model-name">${model}</div>
+                <div class="model-provider">Model Configuration</div>
+            </div>
+            <div class="model-actions">
+                <button class="btn btn-danger" onclick="deleteModel('${model}')">
+                    <i class="fas fa-trash"></i>
+                    Delete
+                </button>
+            </div>
+        </div>
+    `).join('');
+}
+
+// Delete Model
+async function deleteModel(modelName) {
+    if (!confirm(`Are you sure you want to delete the model "${modelName}"?`)) {
+        return;
+    }
+    
+    try {
+        const response = await fetch(`${API_BASE_URL}/models/${modelName}`, {
+            method: 'DELETE'
+        });
+        
+        const result = await response.json();
+        
+        if (response.ok) {
+            showToast('Model deleted successfully!', 'success');
+            loadModels();
+            loadModelSelect();
+        } else {
+            throw new Error(result.detail || 'Failed to delete model');
+        }
+    } catch (error) {
+        console.error('Delete model error:', error);
+        showToast(`Error: ${error.message}`, 'error');
+    }
+}
+
+// Load Model Select Options
+async function loadModelSelect() {
+    try {
+        const response = await fetch(`${API_BASE_URL}/models`);
+        const result = await response.json();
+        
+        if (response.ok) {
+            // Clear existing options
+            modelSelect.innerHTML = '<option value="">Select a Model</option>';
+            
+            // Add model options
+            result.models.forEach(model => {
+                const option = document.createElement('option');
+                option.value = model;
+                option.textContent = model;
+                modelSelect.appendChild(option);
+            });
+        }
+    } catch (error) {
+        console.error('Load model select error:', error);
+    }
+}
+
+// Display Results
+function displayResults(result) {
+    // Update result info
+    document.getElementById('result-url').textContent = result.url;
+    document.getElementById('result-query').textContent = result.query;
+    document.getElementById('result-model').textContent = result.model_name || 'Default Model';
+    
+    // Display JSON in the actual results section
+    const actualJsonOutput = document.getElementById('actual-json-output');
+    actualJsonOutput.textContent = JSON.stringify(result.extracted_data, null, 2);
+    
+    // Don't update the sample JSON in the workflow demo - keep it as example
+    
+    // Update the cURL example based on the approach used
+    const scraping_approach = document.getElementById('scraping-approach').value;
+    const endpoint = scraping_approach === 'llm' ? '/scrape-with-llm' : '/scrape';
+    const curlExample = document.getElementById('curl-example');
+    curlExample.textContent = `curl -X POST http://localhost:8000${endpoint} -H "Content-Type: application/json" -d '{"url": "${result.url}", "query": "${result.query}"}'`;
+    
+    // Show results section
+    resultsSection.style.display = 'block';
+    resultsSection.scrollIntoView({ behavior: 'smooth' });
+}
+
+// Show/Hide Loading
+function showLoading(show) {
+    loadingSection.style.display = show ? 'block' : 'none';
+}
+
+// Hide Results
+function hideResults() {
+    resultsSection.style.display = 'none';
+}
+
+// Toast Notifications
+function showToast(message, type = 'info') {
+    const toastContainer = document.getElementById('toast-container');
+    const toast = document.createElement('div');
+    toast.className = `toast ${type}`;
+    
+    const icon = type === 'success' ? 'fas fa-check-circle' : 
+                 type === 'error' ? 'fas fa-exclamation-circle' : 
+                 'fas fa-info-circle';
+    
+    toast.innerHTML = `
+        <i class="${icon}"></i>
+        <span>${message}</span>
+    `;
+    
+    toastContainer.appendChild(toast);
+    
+    // Auto remove after 5 seconds
+    setTimeout(() => {
+        toast.remove();
+    }, 5000);
+}
+
+// Load Saved Requests
+async function loadSavedRequests() {
+    try {
+        const response = await fetch(`${API_BASE_URL}/saved-requests`);
+        const result = await response.json();
+        
+        if (response.ok) {
+            displaySavedRequests(result.requests);
+        } else {
+            throw new Error(result.detail || 'Failed to load saved requests');
+        }
+    } catch (error) {
+        console.error('Load saved requests error:', error);
+        showToast(`Error: ${error.message}`, 'error');
+    }
+}
+
+// Display Saved Requests
+function displaySavedRequests(requests) {
+    const requestsList = document.getElementById('requests-list');
+    
+    if (requests.length === 0) {
+        requestsList.innerHTML = '<p style="text-align: center; color: #CCCCCC; padding: 2rem;">No saved API requests yet. Make your first request from the Scrape page!</p>';
+        return;
+    }
+    
+    requestsList.innerHTML = requests.map(request => {
+        const url = request.body.url;
+        const query = request.body.query;
+        const model = request.body.model_name || 'Default Model';
+        const endpoint = request.endpoint;
+        
+        // Create curl command
+        const curlCommand = `curl -X POST http://localhost:8000${endpoint} \\
+  -H "Content-Type: application/json" \\
+  -d '{
+    "url": "${url}",
+    "query": "${query}",
+    "model_name": "${model}"
+  }'`;
+        
+        return `
+            <div class="request-card">
+                <div class="request-header">
+                    <div class="request-info">
+                        <div class="request-url">${url}</div>
+                        <div class="request-query">${query}</div>
+                    </div>
+                    <div class="request-actions">
+                        <button class="btn-danger" onclick="deleteSavedRequest('${request.id}')">
+                            <i class="fas fa-trash"></i>
+                            Delete
+                        </button>
+                    </div>
+                </div>
+                
+                <div class="request-curl">
+                    <h4>cURL Command:</h4>
+                    <pre>${curlCommand}</pre>
+                </div>
+            </div>
+        `;
+    }).join('');
+}
+
+// Delete Saved Request
+async function deleteSavedRequest(requestId) {
+    if (!confirm('Are you sure you want to delete this saved request?')) {
+        return;
+    }
+    
+    try {
+        const response = await fetch(`${API_BASE_URL}/saved-requests/${requestId}`, {
+            method: 'DELETE'
+        });
+        
+        const result = await response.json();
+        
+        if (response.ok) {
+            showToast('Saved request deleted successfully!', 'success');
+            loadSavedRequests();
+        } else {
+            throw new Error(result.detail || 'Failed to delete saved request');
+        }
+    } catch (error) {
+        console.error('Delete saved request error:', error);
+        showToast(`Error: ${error.message}`, 'error');
+    }
+}
+
+// Initialize
+document.addEventListener('DOMContentLoaded', () => {
+    loadModelSelect();
+    
+    // Check if API is available
+    fetch(`${API_BASE_URL}/health`)
+        .then(response => {
+            if (!response.ok) {
+                showToast('Warning: API server might not be running', 'error');
+            }
+        })
+        .catch(() => {
+            showToast('Warning: Cannot connect to API server. Make sure it\'s running on localhost:8000', 'error');
+        });
+}); 

docs/examples/website-to-api/static/styles.css

@@ -0,0 +1,765 @@
+/* Reset and Base Styles */
+* {
+    margin: 0;
+    padding: 0;
+    box-sizing: border-box;
+}
+
+body {
+    font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
+    background: #000000;
+    color: #FFFFFF;
+    line-height: 1.6;
+    font-size: 16px;
+}
+
+/* Header */
+.header {
+    border-bottom: 1px solid #333;
+    padding: 1rem 0;
+    background: #000000;
+    position: sticky;
+    top: 0;
+    z-index: 100;
+}
+
+.header-content {
+    max-width: 1200px;
+    margin: 0 auto;
+    padding: 0 2rem;
+    display: flex;
+    justify-content: space-between;
+    align-items: center;
+}
+
+.logo {
+    display: flex;
+    align-items: center;
+    gap: 0.5rem;
+    font-size: 1.5rem;
+    font-weight: 600;
+    color: #FFFFFF;
+}
+
+.logo-image {
+    width: 40px;
+    height: 40px;
+    border-radius: 4px;
+    object-fit: contain;
+}
+
+.nav-links {
+    display: flex;
+    gap: 2rem;
+}
+
+.nav-link {
+    color: #CCCCCC;
+    text-decoration: none;
+    font-weight: 500;
+    transition: color 0.2s ease;
+}
+
+.nav-link:hover,
+.nav-link.active {
+    color: #FFFFFF;
+}
+
+/* Main Content */
+.main-content {
+    max-width: 1200px;
+    margin: 0 auto;
+    padding: 2rem;
+}
+
+.page {
+    display: none;
+}
+
+.page.active {
+    display: block;
+}
+
+/* Hero Section */
+.hero-section {
+    text-align: center;
+    margin-bottom: 4rem;
+    padding: 2rem 0;
+}
+
+.hero-title {
+    font-size: 3rem;
+    font-weight: 700;
+    color: #FFFFFF;
+    margin-bottom: 1rem;
+    line-height: 1.2;
+}
+
+.hero-subtitle {
+    font-size: 1.25rem;
+    color: #CCCCCC;
+    max-width: 600px;
+    margin: 0 auto;
+}
+
+/* Workflow Demo */
+.workflow-demo {
+    display: grid;
+    grid-template-columns: 1fr auto 1fr;
+    gap: 2rem;
+    align-items: start;
+    margin-bottom: 4rem;
+}
+
+.workflow-step {
+    display: flex;
+    flex-direction: column;
+    gap: 1rem;
+}
+
+.step-title {
+    font-size: 1.25rem;
+    font-weight: 600;
+    color: #FFFFFF;
+    text-align: center;
+    margin-bottom: 1rem;
+}
+
+.workflow-arrow {
+    font-size: 2rem;
+    font-weight: 700;
+    color: #09b5a5;
+    display: flex;
+    align-items: center;
+    justify-content: center;
+    margin-top: 20rem;
+}
+
+/* Request Box */
+.request-box {
+    border: 2px solid #333;
+    border-radius: 8px;
+    padding: 2rem;
+    background: #111111;
+}
+
+.input-group {
+    margin-bottom: 1.5rem;
+}
+
+.input-group label {
+    display: block;
+    font-family: 'Courier New', monospace;
+    font-weight: 600;
+    color: #FFFFFF;
+    margin-bottom: 0.5rem;
+    font-size: 0.9rem;
+}
+
+.input-group input,
+.input-group textarea,
+.input-group select {
+    width: 100%;
+    padding: 0.75rem;
+    border: 1px solid #333;
+    border-radius: 4px;
+    font-family: 'Courier New', monospace;
+    font-size: 0.9rem;
+    background: #1A1A1A;
+    color: #FFFFFF;
+    transition: border-color 0.2s ease;
+}
+
+.input-group input:focus,
+.input-group textarea:focus,
+.input-group select:focus {
+    outline: none;
+    border-color: #09b5a5;
+}
+
+.input-group textarea {
+    min-height: 80px;
+    resize: vertical;
+}
+
+.form-options {
+    display: grid;
+    grid-template-columns: 1fr 1fr;
+    gap: 1rem;
+    margin-bottom: 1.5rem;
+}
+
+.option-group {
+    display: flex;
+    flex-direction: column;
+    gap: 0.5rem;
+}
+
+.option-group label {
+    font-family: 'Courier New', monospace;
+    font-weight: 600;
+    color: #FFFFFF;
+    font-size: 0.9rem;
+}
+
+.option-group input[type="checkbox"] {
+    width: auto;
+    margin-right: 0.5rem;
+}
+
+.extract-btn {
+    width: 100%;
+    padding: 1rem;
+    background: #09b5a5;
+    color: #000000;
+    border: none;
+    border-radius: 4px;
+    font-size: 1rem;
+    font-weight: 600;
+    cursor: pointer;
+    transition: background-color 0.2s ease;
+    display: flex;
+    align-items: center;
+    justify-content: center;
+    gap: 0.5rem;
+}
+
+.extract-btn:hover {
+    background: #09b5a5;
+}
+
+/* Dropdown specific styling */
+select,
+.input-group select,
+.option-group select {
+    cursor: pointer !important;
+    appearance: none !important;
+    -webkit-appearance: none !important;
+    -moz-appearance: none !important;
+    -ms-appearance: none !important;
+    background-image: url("data:image/svg+xml;charset=UTF-8,%3csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 24 24' fill='none' stroke='%23FFFFFF' stroke-width='2' stroke-linecap='round' stroke-linejoin='round'%3e%3cpolyline points='6,9 12,15 18,9'%3e%3c/polyline%3e%3c/svg%3e") !important;
+    background-repeat: no-repeat !important;
+    background-position: right 0.75rem center !important;
+    background-size: 1rem !important;
+    padding-right: 2.5rem !important;
+    border: 1px solid #333 !important;
+    border-radius: 4px !important;
+    font-family: 'Courier New', monospace !important;
+    font-size: 0.9rem !important;
+    background-color: #1A1A1A !important;
+    color: #FFFFFF !important;
+}
+
+select:hover,
+.input-group select:hover,
+.option-group select:hover {
+    border-color: #09b5a5 !important;
+}
+
+select:focus,
+.input-group select:focus,
+.option-group select:focus {
+    outline: none !important;
+    border-color: #09b5a5 !important;
+}
+
+select option,
+.input-group select option,
+.option-group select option {
+    background: #1A1A1A !important;
+    color: #FFFFFF !important;
+    padding: 0.5rem !important;
+}
+
+/* Response Container */
+.response-container {
+    display: flex;
+    flex-direction: column;
+    gap: 1rem;
+}
+
+.api-request-box,
+.json-response-box {
+    border: 2px solid #333;
+    border-radius: 8px;
+    padding: 1.5rem;
+    background: #111111;
+}
+
+.api-request-box label,
+.json-response-box label {
+    display: block;
+    font-family: 'Courier New', monospace;
+    font-weight: 600;
+    color: #FFFFFF;
+    margin-bottom: 0.5rem;
+    font-size: 0.9rem;
+}
+
+.api-request-box pre,
+.json-response-box pre {
+    font-family: 'Courier New', monospace;
+    font-size: 0.85rem;
+    line-height: 1.5;
+    color: #FFFFFF;
+    background: #1A1A1A;
+    padding: 1rem;
+    border-radius: 4px;
+    overflow-x: auto;
+    white-space: pre-wrap;
+    word-break: break-all;
+}
+
+/* Results Section */
+.results-section {
+    border: 2px solid #333;
+    border-radius: 8px;
+    overflow: hidden;
+    margin-top: 2rem;
+    background: #111111;
+}
+
+.results-header {
+    background: #1A1A1A;
+    color: #FFFFFF;
+    padding: 1rem 1.5rem;
+    display: flex;
+    justify-content: space-between;
+    align-items: center;
+    border-bottom: 1px solid #333;
+}
+
+.results-header h2 {
+    font-size: 1.25rem;
+    font-weight: 600;
+    color: #FFFFFF;
+}
+
+.copy-btn {
+    background: #09b5a5;
+    color: #000000;
+    border: none;
+    padding: 0.5rem 1rem;
+    border-radius: 4px;
+    font-size: 0.9rem;
+    font-weight: 600;
+    cursor: pointer;
+    display: flex;
+    align-items: center;
+    gap: 0.5rem;
+    transition: background-color 0.2s ease;
+}
+
+.copy-btn:hover {
+    background: #09b5a5;
+}
+
+.results-content {
+    padding: 1.5rem;
+}
+
+.result-info {
+    display: grid;
+    grid-template-columns: repeat(auto-fit, minmax(250px, 1fr));
+    gap: 1rem;
+    margin-bottom: 1.5rem;
+    padding: 1rem;
+    background: #1A1A1A;
+    border-radius: 4px;
+    border: 1px solid #333;
+}
+
+.info-item {
+    display: flex;
+    flex-direction: column;
+    gap: 0.25rem;
+}
+
+.info-item .label {
+    font-weight: 600;
+    color: #FFFFFF;
+    font-size: 0.9rem;
+}
+
+.info-item .value {
+    color: #CCCCCC;
+    word-break: break-all;
+}
+
+.json-display {
+    background: #1A1A1A;
+    border-radius: 4px;
+    overflow: hidden;
+    border: 1px solid #333;
+}
+
+.json-display pre {
+    color: #FFFFFF;
+    padding: 1.5rem;
+    margin: 0;
+    overflow-x: auto;
+    font-family: 'Courier New', monospace;
+    font-size: 0.9rem;
+    line-height: 1.5;
+}
+
+/* Loading State */
+.loading {
+    text-align: center;
+    padding: 3rem;
+}
+
+.spinner {
+    width: 40px;
+    height: 40px;
+    border: 3px solid #333;
+    border-top: 3px solid #09b5a5;
+    border-radius: 50%;
+    animation: spin 1s linear infinite;
+    margin: 0 auto 1rem;
+}
+
+@keyframes spin {
+    0% { transform: rotate(0deg); }
+    100% { transform: rotate(360deg); }
+}
+
+/* Models Page */
+.models-header {
+    text-align: center;
+    margin-bottom: 3rem;
+}
+
+.models-header h1 {
+    font-size: 2.5rem;
+    font-weight: 700;
+    color: #FFFFFF;
+    margin-bottom: 1rem;
+}
+
+.models-header p {
+    font-size: 1.1rem;
+    color: #CCCCCC;
+}
+
+/* API Requests Page */
+.requests-header {
+    text-align: center;
+    margin-bottom: 3rem;
+}
+
+.requests-header h1 {
+    font-size: 2.5rem;
+    font-weight: 700;
+    color: #FFFFFF;
+    margin-bottom: 1rem;
+}
+
+.requests-header p {
+    font-size: 1.1rem;
+    color: #CCCCCC;
+}
+
+.requests-container {
+    max-width: 1200px;
+    margin: 0 auto;
+}
+
+.requests-list {
+    display: grid;
+    gap: 1.5rem;
+}
+
+.request-card {
+    border: 2px solid #333;
+    border-radius: 8px;
+    padding: 1.5rem;
+    background: #111111;
+    transition: border-color 0.2s ease;
+}
+
+.request-card:hover {
+    border-color: #09b5a5;
+}
+
+.request-header {
+    display: flex;
+    justify-content: space-between;
+    align-items: center;
+    margin-bottom: 1rem;
+    padding-bottom: 1rem;
+    border-bottom: 1px solid #333;
+}
+
+.request-info {
+    display: flex;
+    flex-direction: column;
+    gap: 0.5rem;
+}
+
+.request-url {
+    font-family: 'Courier New', monospace;
+    font-weight: 600;
+    color: #09b5a5;
+    font-size: 1.1rem;
+    word-break: break-all;
+}
+
+.request-query {
+    color: #CCCCCC;
+    font-size: 0.9rem;
+    margin-top: 0.5rem;
+    word-break: break-all;
+}
+
+.request-actions {
+    display: flex;
+    gap: 0.5rem;
+}
+
+.request-curl {
+    background: #1A1A1A;
+    border: 1px solid #333;
+    border-radius: 4px;
+    padding: 1rem;
+    margin-top: 1rem;
+}
+
+.request-curl h4 {
+    color: #FFFFFF;
+    font-size: 0.9rem;
+    font-weight: 600;
+    margin-bottom: 0.5rem;
+    font-family: 'Courier New', monospace;
+}
+
+.request-curl pre {
+    color: #CCCCCC;
+    font-size: 0.8rem;
+    line-height: 1.4;
+    overflow-x: auto;
+    white-space: pre-wrap;
+    word-break: break-all;
+    background: #111111;
+    padding: 0.75rem;
+    border-radius: 4px;
+    border: 1px solid #333;
+}
+
+.models-container {
+    max-width: 800px;
+    margin: 0 auto;
+}
+
+.model-form-section {
+    border: 2px solid #333;
+    border-radius: 8px;
+    padding: 2rem;
+    margin-bottom: 2rem;
+    background: #111111;
+}
+
+.model-form-section h3 {
+    font-size: 1.25rem;
+    font-weight: 600;
+    color: #FFFFFF;
+    margin-bottom: 1.5rem;
+}
+
+.model-form {
+    display: flex;
+    flex-direction: column;
+    gap: 1.5rem;
+}
+
+.form-row {
+    display: grid;
+    grid-template-columns: 1fr 1fr;
+    gap: 1rem;
+}
+
+.save-btn {
+    padding: 1rem;
+    background: #09b5a5;
+    color: #000000;
+    border: none;
+    border-radius: 4px;
+    font-size: 1rem;
+    font-weight: 600;
+    cursor: pointer;
+    transition: background-color 0.2s ease;
+    display: flex;
+    align-items: center;
+    justify-content: center;
+    gap: 0.5rem;
+}
+
+.save-btn:hover {
+    background: #09b5a5;
+}
+
+.saved-models-section h3 {
+    font-size: 1.25rem;
+    font-weight: 600;
+    color: #FFFFFF;
+    margin-bottom: 1.5rem;
+}
+
+.models-list {
+    display: grid;
+    gap: 1rem;
+}
+
+.model-card {
+    border: 2px solid #333;
+    border-radius: 8px;
+    padding: 1.5rem;
+    display: flex;
+    justify-content: space-between;
+    align-items: center;
+    transition: border-color 0.2s ease;
+    background: #111111;
+}
+
+.model-card:hover {
+    border-color: #09b5a5;
+}
+
+.model-info {
+    flex: 1;
+}
+
+.model-name {
+    font-weight: 600;
+    color: #FFFFFF;
+    font-size: 1.1rem;
+    margin-bottom: 0.5rem;
+}
+
+.model-provider {
+    color: #CCCCCC;
+    font-size: 0.9rem;
+}
+
+.model-actions {
+    display: flex;
+    gap: 0.5rem;
+}
+
+.btn-danger {
+    background: #FF4444;
+    color: #FFFFFF;
+    border: none;
+    padding: 0.5rem 1rem;
+    border-radius: 4px;
+    font-size: 0.9rem;
+    font-weight: 600;
+    cursor: pointer;
+    transition: background-color 0.2s ease;
+    display: flex;
+    align-items: center;
+    gap: 0.5rem;
+}
+
+.btn-danger:hover {
+    background: #CC3333;
+}
+
+
+
+/* Toast Notifications */
+.toast-container {
+    position: fixed;
+    top: 20px;
+    right: 20px;
+    z-index: 1000;
+}
+
+.toast {
+    background: #111111;
+    border: 2px solid #333;
+    border-radius: 4px;
+    padding: 1rem 1.5rem;
+    margin-bottom: 0.5rem;
+    display: flex;
+    align-items: center;
+    gap: 0.5rem;
+    animation: slideIn 0.3s ease;
+    max-width: 400px;
+    box-shadow: 0 4px 12px rgba(0, 0, 0, 0.3);
+    color: #FFFFFF;
+}
+
+.toast.success {
+    border-color: #09b5a5;
+    background: #0A1A1A;
+}
+
+.toast.error {
+    border-color: #FF4444;
+    background: #1A0A0A;
+}
+
+.toast.info {
+    border-color: #09b5a5;
+    background: #0A1A1A;
+}
+
+@keyframes slideIn {
+    from {
+        transform: translateX(100%);
+        opacity: 0;
+    }
+    to {
+        transform: translateX(0);
+        opacity: 1;
+    }
+}
+
+/* Responsive Design */
+@media (max-width: 768px) {
+    .header-content {
+        padding: 0 1rem;
+    }
+    
+    .main-content {
+        padding: 1rem;
+    }
+    
+    .hero-title {
+        font-size: 2rem;
+    }
+    
+    .workflow-demo {
+        grid-template-columns: 1fr;
+        gap: 1rem;
+    }
+    
+    .workflow-arrow {
+        transform: rotate(90deg);
+        margin: 1rem 0;
+    }
+    
+    .form-options {
+        grid-template-columns: 1fr;
+    }
+    
+    .form-row {
+        grid-template-columns: 1fr;
+    }
+    
+    .result-info {
+        grid-template-columns: 1fr;
+    }
+    
+    .model-card {
+        flex-direction: column;
+        gap: 1rem;
+        text-align: center;
+    }
+    
+    .model-actions {
+        width: 100%;
+        justify-content: center;
+    }
+}

docs/examples/website-to-api/test_api.py

@@ -0,0 +1,28 @@
+import asyncio
+from web_scraper_lib import scrape_website
+import os
+
+async def test_library():
+    """Test the mini library directly."""
+    print("=== Testing Mini Library ===")
+    
+    # Test 1: Scrape with a custom model
+    url = "https://marketplace.mainstreet.co.in/collections/adidas-yeezy/products/adidas-yeezy-boost-350-v2-yecheil-non-reflective"
+    query = "Extract the following data: Product name, Product price, Product description, Product size. DO NOT EXTRACT ANYTHING ELSE."
+    if os.path.exists("models"):
+        model_name = os.listdir("models")[0].split(".")[0]
+    else:
+        raise Exception("No models found in models directory")
+
+    print(f"Scraping: {url}")
+    print(f"Query: {query}")
+    
+    try:
+        result = await scrape_website(url, query, model_name)
+        print("✅ Library test successful!")
+        print(f"Extracted data: {result['extracted_data']}")
+    except Exception as e:
+        print(f"❌ Library test failed: {e}")
+
+if __name__ == "__main__":
+    asyncio.run(test_library())

docs/examples/website-to-api/test_models.py

@@ -0,0 +1,67 @@
+#!/usr/bin/env python3
+"""
+Test script for the new model management functionality.
+This script demonstrates how to save and use custom model configurations.
+"""
+
+import asyncio
+import requests
+import json
+
+# API base URL
+BASE_URL = "http://localhost:8000"
+
+def test_model_management():
+    """Test the model management endpoints."""
+    
+    print("=== Testing Model Management ===")
+    
+    # 1. List current models
+    print("\n1. Listing current models:")
+    response = requests.get(f"{BASE_URL}/models")
+    print(f"Status: {response.status_code}")
+    print(f"Response: {json.dumps(response.json(), indent=2)}")
+    
+    
+    # 2. Save another model configuration (OpenAI example)
+    print("\n2. Saving OpenAI model configuration:")
+    openai_config = {
+        "model_name": "my-openai",
+        "provider": "openai",
+        "api_token": "your-openai-api-key-here"
+    }
+    
+    response = requests.post(f"{BASE_URL}/models", json=openai_config)
+    print(f"Status: {response.status_code}")
+    print(f"Response: {json.dumps(response.json(), indent=2)}")
+    
+    # 3. List models again to see the new ones
+    print("\n3. Listing models after adding new ones:")
+    response = requests.get(f"{BASE_URL}/models")
+    print(f"Status: {response.status_code}")
+    print(f"Response: {json.dumps(response.json(), indent=2)}")
+
+    # 4. Delete a model configuration
+    print("\n4. Deleting a model configuration:")
+    response = requests.delete(f"{BASE_URL}/models/my-openai")
+    print(f"Status: {response.status_code}")
+    print(f"Response: {json.dumps(response.json(), indent=2)}")
+    
+    # 5. Final list of models
+    print("\n5. Final list of models:")
+    response = requests.get(f"{BASE_URL}/models")
+    print(f"Status: {response.status_code}")
+    print(f"Response: {json.dumps(response.json(), indent=2)}")
+
+if __name__ == "__main__":
+    print("Model Management Test Script")
+    print("Make sure the API server is running on http://localhost:8000")
+    print("=" * 50)
+    
+    try:
+        test_model_management()
+    except requests.exceptions.ConnectionError:
+        print("Error: Could not connect to the API server.")
+        print("Make sure the server is running with: python api_server.py")
+    except Exception as e:
+        print(f"Error: {e}") 

docs/examples/website-to-api/web_scraper_lib.py

@@ -0,0 +1,397 @@
+from crawl4ai import (
+    AsyncWebCrawler,
+    BrowserConfig,
+    CacheMode,
+    CrawlerRunConfig,
+    LLMConfig,
+    JsonCssExtractionStrategy,
+    LLMExtractionStrategy
+)
+import os
+import json
+import hashlib
+from typing import Dict, Any, Optional, List
+from litellm import completion
+
+class ModelConfig:
+    """Configuration for LLM models."""
+    
+    def __init__(self, provider: str, api_token: str):
+        self.provider = provider
+        self.api_token = api_token
+    
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "provider": self.provider,
+            "api_token": self.api_token
+        }
+    
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> 'ModelConfig':
+        return cls(
+            provider=data["provider"],
+            api_token=data["api_token"]
+        )
+
+class WebScraperAgent:
+    """
+    A mini library that converts any website into a structured data API.
+    
+    Features:
+    1. Provide a URL and tell AI what data you need in plain English
+    2. Generate: Agent reverse-engineers the site and deploys custom scraper
+    3. Integrate: Use private API endpoint to get structured data
+    4. Support for custom LLM models and API keys
+    """
+    
+    def __init__(self, schemas_dir: str = "schemas", models_dir: str = "models"):
+        self.schemas_dir = schemas_dir
+        self.models_dir = models_dir
+        os.makedirs(self.schemas_dir, exist_ok=True)
+        os.makedirs(self.models_dir, exist_ok=True)
+    
+    def _generate_schema_key(self, url: str, query: str) -> str:
+        """Generate a unique key for schema caching based on URL and query."""
+        content = f"{url}:{query}"
+        return hashlib.md5(content.encode()).hexdigest()
+    
+    def save_model_config(self, model_name: str, provider: str, api_token: str) -> bool:
+        """
+        Save a model configuration for later use.
+        
+        Args:
+            model_name: User-friendly name for the model
+            provider: LLM provider (e.g., 'gemini', 'openai', 'anthropic')
+            api_token: API token for the provider
+            
+        Returns:
+            True if saved successfully
+        """
+        try:
+            model_config = ModelConfig(provider, api_token)
+            config_path = os.path.join(self.models_dir, f"{model_name}.json")
+            
+            with open(config_path, "w") as f:
+                json.dump(model_config.to_dict(), f, indent=2)
+            
+            print(f"Model configuration saved: {model_name}")
+            return True
+        except Exception as e:
+            print(f"Failed to save model configuration: {e}")
+            return False
+    
+    def load_model_config(self, model_name: str) -> Optional[ModelConfig]:
+        """
+        Load a saved model configuration.
+        
+        Args:
+            model_name: Name of the saved model configuration
+            
+        Returns:
+            ModelConfig object or None if not found
+        """
+        try:
+            config_path = os.path.join(self.models_dir, f"{model_name}.json")
+            if not os.path.exists(config_path):
+                return None
+            
+            with open(config_path, "r") as f:
+                data = json.load(f)
+            
+            return ModelConfig.from_dict(data)
+        except Exception as e:
+            print(f"Failed to load model configuration: {e}")
+            return None
+    
+    def list_saved_models(self) -> List[str]:
+        """List all saved model configurations."""
+        models = []
+        for filename in os.listdir(self.models_dir):
+            if filename.endswith('.json'):
+                models.append(filename[:-5])  # Remove .json extension
+        return models
+    
+    def delete_model_config(self, model_name: str) -> bool:
+        """
+        Delete a saved model configuration.
+        
+        Args:
+            model_name: Name of the model configuration to delete
+            
+        Returns:
+            True if deleted successfully
+        """
+        try:
+            config_path = os.path.join(self.models_dir, f"{model_name}.json")
+            if os.path.exists(config_path):
+                os.remove(config_path)
+                print(f"Model configuration deleted: {model_name}")
+                return True
+            return False
+        except Exception as e:
+            print(f"Failed to delete model configuration: {e}")
+            return False
+    
+    async def _load_or_generate_schema(self, url: str, query: str, session_id: str = "schema_generator", model_name: Optional[str] = None) -> Dict[str, Any]:
+        """
+        Loads schema from cache if exists, otherwise generates using AI.
+        This is the "Generate" step - our agent reverse-engineers the site.
+        
+        Args:
+            url: URL to scrape
+            query: Query for data extraction
+            session_id: Session identifier
+            model_name: Name of saved model configuration to use
+        """
+        schema_key = self._generate_schema_key(url, query)
+        schema_path = os.path.join(self.schemas_dir, f"{schema_key}.json")
+        
+        if os.path.exists(schema_path):
+            print(f"Schema found in cache for {url}")
+            with open(schema_path, "r") as f:
+                return json.load(f)
+        
+        print(f"Generating new schema for {url}")
+        print(f"Query: {query}")
+        query += """
+        IMPORTANT:
+        GENERATE THE SCHEMA WITH ONLY THE FIELDS MENTIONED IN THE QUERY. MAKE SURE THE NUMBER OF FIELDS IN THE SCHEME MATCH THE NUMBER OF FIELDS IN THE QUERY.
+        """
+        
+        # Step 1: Fetch the page HTML
+        async with AsyncWebCrawler(config=BrowserConfig(headless=True)) as crawler:
+            result = await crawler.arun(
+                url=url,
+                config=CrawlerRunConfig(
+                    cache_mode=CacheMode.BYPASS,
+                    session_id=session_id,
+                    simulate_user=True,
+                    remove_overlay_elements=True,
+                    delay_before_return_html=5,
+                )
+            )
+            html = result.fit_html
+        
+        # Step 2: Generate schema using AI with custom model if specified
+        print("AI is analyzing the page structure...")
+        
+        # Use custom model configuration if provided
+        if model_name:
+            model_config = self.load_model_config(model_name)
+            if model_config:
+                llm_config = LLMConfig(
+                    provider=model_config.provider,
+                    api_token=model_config.api_token
+                )
+                print(f"Using custom model: {model_name}")
+            else:
+                raise ValueError(f"Model configuration '{model_name}' not found. Please add it from the Models page.")
+        else:
+            # Require a model to be specified
+            raise ValueError("No model specified. Please select a model from the dropdown or add one from the Models page.")
+        
+        schema = JsonCssExtractionStrategy.generate_schema(
+            html=html,
+            llm_config=llm_config,
+            query=query
+        )
+        
+        # Step 3: Cache the generated schema
+        print(f"Schema generated and cached: {json.dumps(schema, indent=2)}")
+        with open(schema_path, "w") as f:
+            json.dump(schema, f, indent=2)
+        
+        return schema
+    
+    def _generate_llm_schema(self, query: str, llm_config: LLMConfig) -> Dict[str, Any]:
+        """
+        Generate a schema for a given query using a custom LLM model.
+        
+        Args:
+            query: Plain English description of what data to extract
+            model_config: Model configuration to use
+        """
+        # ask the model to generate a schema for the given query in the form of a json.
+        prompt = f"""
+        IDENTIFY THE FIELDS FOR EXTRACTION MENTIONED IN THE QUERY and GENERATE A JSON SCHEMA FOR THE FIELDS.
+        eg.
+        {{
+            "name": "str",  
+            "age": "str",
+            "email": "str",
+            "product_name": "str",
+            "product_price": "str",
+            "product_description": "str",
+            "product_image": "str",
+            "product_url": "str",
+            "product_rating": "str",
+            "product_reviews": "str",
+        }}
+        Here is the query:
+        {query}
+        IMPORTANT:
+        THE RESULT SHOULD BE A JSON OBJECT.
+        MAKE SURE THE NUMBER OF FIELDS IN THE RESULT MATCH THE NUMBER OF FIELDS IN THE QUERY.
+        THE RESULT SHOULD BE A JSON OBJECT.
+        """
+        response = completion(
+            model=llm_config.provider,
+            messages=[{"role": "user", "content": prompt}],
+            api_key=llm_config.api_token,
+            result_type="json"
+        )
+
+        return response.json()["choices"][0]["message"]["content"]
+    async def scrape_data_with_llm(self, url: str, query: str, model_name: Optional[str] = None) -> Dict[str, Any]:
+        """
+        Scrape structured data from any website using a custom LLM model.
+        
+        Args:
+            url: The website URL to scrape
+            query: Plain English description of what data to extract
+            model_name: Name of saved model configuration to use
+        """
+
+        if model_name:
+            model_config = self.load_model_config(model_name)
+            if model_config:
+                llm_config = LLMConfig(
+                    provider=model_config.provider,
+                    api_token=model_config.api_token
+                )
+                print(f"Using custom model: {model_name}")
+            else:
+                raise ValueError(f"Model configuration '{model_name}' not found. Please add it from the Models page.")
+        else:
+            # Require a model to be specified
+            raise ValueError("No model specified. Please select a model from the dropdown or add one from the Models page.")
+
+        query += """\n  
+        IMPORTANT:
+        THE RESULT SHOULD BE A JSON OBJECT WITH THE ONLY THE FIELDS MENTIONED IN THE QUERY.
+        MAKE SURE THE NUMBER OF FIELDS IN THE RESULT MATCH THE NUMBER OF FIELDS IN THE QUERY.
+        THE RESULT SHOULD BE A JSON OBJECT.
+        """
+
+        schema = self._generate_llm_schema(query, llm_config)
+
+        print(f"Schema: {schema}")
+
+        llm_extraction_strategy = LLMExtractionStrategy(
+            llm_config=llm_config,
+            instruction=query,
+            result_type="json",
+            schema=schema
+        )
+
+        async with AsyncWebCrawler() as crawler:
+            result = await crawler.arun(
+                url=url,
+                config=CrawlerRunConfig(
+                    cache_mode=CacheMode.BYPASS,
+                    simulate_user=True,
+                    extraction_strategy=llm_extraction_strategy,
+                )
+            )
+        extracted_data = result.extracted_content
+        if isinstance(extracted_data, str):
+                try:
+                    extracted_data = json.loads(extracted_data)
+                except json.JSONDecodeError:
+                    # If it's not valid JSON, keep it as string
+                    pass
+            
+        return {
+                "url": url,
+                "query": query,
+                "extracted_data": extracted_data,
+                "timestamp": result.timestamp if hasattr(result, 'timestamp') else None
+            }
+        
+    async def scrape_data(self, url: str, query: str, model_name: Optional[str] = None) -> Dict[str, Any]:
+        """
+        Main method to scrape structured data from any website.
+        
+        Args:
+            url: The website URL to scrape
+            query: Plain English description of what data to extract
+            model_name: Name of saved model configuration to use
+            
+        Returns:
+            Structured data extracted from the website
+        """
+        # Step 1: Generate or load schema (reverse-engineer the site)
+        schema = await self._load_or_generate_schema(url=url, query=query, model_name=model_name)
+        
+        # Step 2: Deploy custom high-speed scraper
+        print(f"Deploying custom scraper for {url}")
+        browser_config = BrowserConfig(headless=True)
+        
+        async with AsyncWebCrawler(config=browser_config) as crawler:
+            run_config = CrawlerRunConfig(
+                extraction_strategy=JsonCssExtractionStrategy(schema=schema),
+            )
+            result = await crawler.arun(url=url, config=run_config)
+            
+            # Step 3: Return structured data
+            # Parse extracted_content if it's a JSON string
+            extracted_data = result.extracted_content
+            if isinstance(extracted_data, str):
+                try:
+                    extracted_data = json.loads(extracted_data)
+                except json.JSONDecodeError:
+                    # If it's not valid JSON, keep it as string
+                    pass
+            
+            return {
+                "url": url,
+                "query": query,
+                "extracted_data": extracted_data,
+                "schema_used": schema,
+                "timestamp": result.timestamp if hasattr(result, 'timestamp') else None
+            }
+    
+    async def get_cached_schemas(self) -> Dict[str, str]:
+        """Get list of cached schemas."""
+        schemas = {}
+        for filename in os.listdir(self.schemas_dir):
+            if filename.endswith('.json'):
+                schema_key = filename[:-5]  # Remove .json extension
+                schemas[schema_key] = filename
+        return schemas
+    
+    def clear_cache(self):
+        """Clear all cached schemas."""
+        import shutil
+        if os.path.exists(self.schemas_dir):
+            shutil.rmtree(self.schemas_dir)
+        os.makedirs(self.schemas_dir, exist_ok=True)
+        print("Schema cache cleared")
+
+# Convenience function for simple usage
+async def scrape_website(url: str, query: str, model_name: Optional[str] = None) -> Dict[str, Any]:
+    """
+    Simple function to scrape any website with plain English instructions.
+    
+    Args:
+        url: Website URL
+        query: Plain English description of what data to extract
+        model_name: Name of saved model configuration to use
+        
+    Returns:
+        Extracted structured data
+    """
+    agent = WebScraperAgent()
+    return await agent.scrape_data(url, query, model_name)
+
+async def scrape_website_with_llm(url: str, query: str, model_name: Optional[str] = None):
+    """
+    Scrape structured data from any website using a custom LLM model.
+    
+    Args:
+        url: The website URL to scrape
+        query: Plain English description of what data to extract
+        model_name: Name of saved model configuration to use
+    """
+    agent = WebScraperAgent()
+    return await agent.scrape_data_with_llm(url, query, model_name)

docs/md_v2/advanced/adaptive-strategies.md

@@ -126,30 +126,6 @@ Factors:
 - URL depth (fewer slashes = higher authority)
 - Clean URL structure
 
-### Custom Link Scoring
-
-```python
-class CustomLinkScorer:
-    def score(self, link: Link, query: str, state: CrawlState) -> float:
-        # Prioritize specific URL patterns
-        if "/api/reference/" in link.href:
-            return 2.0  # Double the score
-        
-        # Deprioritize certain sections
-        if "/archive/" in link.href:
-            return 0.1  # Reduce score by 90%
-        
-        # Default scoring
-        return 1.0
-
-# Use with adaptive crawler
-adaptive = AdaptiveCrawler(
-    crawler,
-    config=config,
-    link_scorer=CustomLinkScorer()
-)
-```
-
 ## Domain-Specific Configurations
 
 ### Technical Documentation
@@ -230,8 +206,12 @@ config = AdaptiveConfig(
 
 # Periodically clean state
 if len(state.knowledge_base) > 1000:
-    # Keep only most relevant
-    state.knowledge_base = get_top_relevant(state.knowledge_base, 500)
+    # Keep only the top 500 most relevant docs
+    top_content = adaptive.get_relevant_content(top_k=500)
+    keep_indices = {d["index"] for d in top_content}
+    state.knowledge_base = [
+        doc for i, doc in enumerate(state.knowledge_base) if i in keep_indices
+    ]
 ```
 
 ### Parallel Processing
@@ -252,18 +232,6 @@ tasks = [
 results = await asyncio.gather(*tasks)
 ```
 
-### Caching Strategy
-
-```python
-# Enable caching for repeated crawls
-async with AsyncWebCrawler(
-    config=BrowserConfig(
-        cache_mode=CacheMode.ENABLED
-    )
-) as crawler:
-    adaptive = AdaptiveCrawler(crawler, config)
-```
-
 ## Debugging & Analysis
 
 ### Enable Verbose Logging
@@ -322,9 +290,9 @@ with open("crawl_analysis.json", "w") as f:
 ### Implementing a Custom Strategy
 
 ```python
-from crawl4ai.adaptive_crawler import BaseStrategy
+from crawl4ai.adaptive_crawler import CrawlStrategy
 
-class DomainSpecificStrategy(BaseStrategy):
+class DomainSpecificStrategy(CrawlStrategy):
     def calculate_coverage(self, state: CrawlState) -> float:
         # Custom coverage calculation
         # e.g., weight certain terms more heavily
@@ -351,7 +319,7 @@ adaptive = AdaptiveCrawler(
 ### Combining Strategies
 
 ```python
-class HybridStrategy(BaseStrategy):
+class HybridStrategy(CrawlStrategy):
     def __init__(self):
         self.strategies = [
             TechnicalDocStrategy(),

docs/md_v2/api/parameters.md

@@ -155,6 +155,7 @@ If your page is a single-page app with repeated JS updates, set `js_only=True` i
 | **`exclude_external_links`** | `bool` (False)          | Removes all links pointing outside the current domain.                                                                      |
 | **`exclude_social_media_links`** | `bool` (False)      | Strips links specifically to social sites (like Facebook or Twitter).                                                      |
 | **`exclude_domains`**        | `list` ([])             | Provide a custom list of domains to exclude (like `["ads.com", "trackers.io"]`).                                            |
+| **`preserve_https_for_internal_links`** | `bool` (False) | If `True`, preserves HTTPS scheme for internal links even when the server redirects to HTTP. Useful for security-conscious crawling. |
 
 Use these for link-level content filtering (often to keep crawls “internal” or to remove spammy domains).
 

docs/md_v2/core/deep-crawling.md

@@ -472,6 +472,17 @@ Note that for BestFirstCrawlingStrategy, score_threshold is not needed since pag
 
 5.**Balance breadth vs. depth.** Choose your strategy wisely - BFS for comprehensive coverage, DFS for deep exploration, BestFirst for focused relevance-based crawling.
 
+6.**Preserve HTTPS for security.** If crawling HTTPS sites that redirect to HTTP, use `preserve_https_for_internal_links=True` to maintain secure connections:
+
+```python
+config = CrawlerRunConfig(
+    deep_crawl_strategy=BFSDeepCrawlStrategy(max_depth=2),
+    preserve_https_for_internal_links=True  # Keep HTTPS even if server redirects to HTTP
+)
+```
+
+This is especially useful for security-conscious crawling or when dealing with sites that support both protocols.
+
 ---
 
 ## 10. Summary & Next Steps

docs/md_v2/core/quickstart.md

@@ -79,7 +79,7 @@ if __name__ == "__main__":
     asyncio.run(main())
 ```
 
-> IMPORTANT: By default cache mode is set to `CacheMode.ENABLED`. So to have fresh content, you need to set it to `CacheMode.BYPASS`
+> IMPORTANT: By default cache mode is set to `CacheMode.BYPASS` to have fresh content. Set `CacheMode.ENABLED` to enable caching.
 
 We’ll explore more advanced config in later tutorials (like enabling proxies, PDF output, multi-tab sessions, etc.). For now, just note how you pass these objects to manage crawling.
 

pyproject.toml

@@ -7,7 +7,7 @@ name = "Crawl4AI"
 dynamic = ["version"]
 description = "🚀🤖 Crawl4AI: Open-source LLM Friendly Web Crawler & scraper"
 readme = "README.md"
-requires-python = ">=3.9"
+requires-python = ">=3.10"
 license = "Apache-2.0"
 authors = [
     {name = "Unclecode", email = "unclecode@kidocode.com"}
@@ -36,6 +36,7 @@ dependencies = [
     "PyYAML>=6.0",
     "nltk>=3.9.1",
     "rich>=13.9.4",
+    "cssselect>=1.2.0",
     "httpx>=0.27.2",
     "httpx[http2]>=0.27.2",
     "fake-useragent>=2.0.3",
@@ -51,7 +52,6 @@ classifiers = [
     "Development Status :: 4 - Beta",
     "Intended Audience :: Developers",
     "Programming Language :: Python :: 3",
-    "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",

requirements.txt

@@ -24,6 +24,7 @@ psutil>=6.1.1
 PyYAML>=6.0
 nltk>=3.9.1
 rich>=13.9.4
+cssselect>=1.2.0
 chardet>=5.2.0
 brotli>=1.1.0
 httpx[http2]>=0.27.2

setup.py

@@ -56,11 +56,10 @@ setup(
         "Development Status :: 3 - Alpha",
         "Intended Audience :: Developers",
         "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.9",
         "Programming Language :: Python :: 3.10",
         "Programming Language :: Python :: 3.11",
         "Programming Language :: Python :: 3.12",
         "Programming Language :: Python :: 3.13",
     ],
-    python_requires=">=3.9",
+    python_requires=">=3.10",
 )

tests/docker/test_filter_deep_crawl.py

@@ -0,0 +1,201 @@
+"""
+Test the complete fix for both the filter serialization and JSON serialization issues.
+"""
+
+import asyncio
+import httpx
+
+from crawl4ai import BrowserConfig, CacheMode, CrawlerRunConfig
+from crawl4ai.deep_crawling import BFSDeepCrawlStrategy, FilterChain, URLPatternFilter
+
+BASE_URL = "http://localhost:11234/"  # Adjust port as needed
+
+async def test_with_docker_client():
+    """Test using the Docker client (same as 1419.py)."""
+    from crawl4ai.docker_client import Crawl4aiDockerClient
+    
+    print("=" * 60)
+    print("Testing with Docker Client")
+    print("=" * 60)
+    
+    try:
+        async with Crawl4aiDockerClient(
+            base_url=BASE_URL,
+            verbose=True,
+        ) as client:
+            
+            # Create filter chain - testing the serialization fix
+            filter_chain = [
+                URLPatternFilter(
+                    # patterns=["*about*", "*privacy*", "*terms*"],
+                    patterns=["*advanced*"],
+                    reverse=True
+                ),
+            ]
+            
+            crawler_config = CrawlerRunConfig(
+                deep_crawl_strategy=BFSDeepCrawlStrategy(
+                    max_depth=2,  # Keep it shallow for testing
+                    # max_pages=5,  # Limit pages for testing
+                    filter_chain=FilterChain(filter_chain)
+                ),
+                cache_mode=CacheMode.BYPASS,
+            )
+            
+            print("\n1. Testing crawl with filters...")
+            results = await client.crawl(
+                ["https://docs.crawl4ai.com"],  # Simple test page
+                browser_config=BrowserConfig(headless=True),
+                crawler_config=crawler_config,
+            )
+            
+            if results:
+                print(f"✅ Crawl succeeded! Type: {type(results)}")
+                if hasattr(results, 'success'):
+                    print(f"✅ Results success: {results.success}")
+                    # Test that we can iterate results without JSON errors
+                    if hasattr(results, '__iter__'):
+                        for i, result in enumerate(results):
+                            if hasattr(result, 'url'):
+                                print(f"   Result {i}: {result.url[:50]}...")
+                            else:
+                                print(f"   Result {i}: {str(result)[:50]}...")
+                else:
+                    # Handle list of results
+                    print(f"✅ Got {len(results)} results")
+                    for i, result in enumerate(results[:3]):  # Show first 3
+                        print(f"   Result {i}: {result.url[:50]}...")
+            else:
+                print("❌ Crawl failed - no results returned")
+                return False
+                
+        print("\n✅ Docker client test completed successfully!")
+        return True
+        
+    except Exception as e:
+        print(f"❌ Docker client test failed: {e}")
+        import traceback
+        traceback.print_exc()
+        return False
+
+
+async def test_with_rest_api():
+    """Test using REST API directly."""
+    print("\n" + "=" * 60)
+    print("Testing with REST API")
+    print("=" * 60)
+    
+    # Create filter configuration
+    deep_crawl_strategy_payload = {
+        "type": "BFSDeepCrawlStrategy",
+        "params": {
+            "max_depth": 2,
+            # "max_pages": 5,
+            "filter_chain": {
+                "type": "FilterChain",
+                "params": {
+                    "filters": [
+                        {
+                            "type": "URLPatternFilter",
+                            "params": {
+                                "patterns": ["*advanced*"],
+                                "reverse": True
+                            }
+                        }
+                    ]
+                }
+            }
+        }
+    }
+    
+    crawl_payload = {
+        "urls": ["https://docs.crawl4ai.com"],
+        "browser_config": {"type": "BrowserConfig", "params": {"headless": True}},
+        "crawler_config": {
+            "type": "CrawlerRunConfig",
+            "params": {
+                "deep_crawl_strategy": deep_crawl_strategy_payload,
+                "cache_mode": "bypass"
+            }
+        }
+    }
+    
+    try:
+        async with httpx.AsyncClient() as client:
+            print("\n1. Sending crawl request to REST API...")
+            response = await client.post(
+                f"{BASE_URL}crawl",
+                json=crawl_payload,
+                timeout=30
+            )
+            
+            if response.status_code == 200:
+                print(f"✅ REST API returned 200 OK")
+                data = response.json()
+                if data.get("success"):
+                    results = data.get("results", [])
+                    print(f"✅ Got {len(results)} results")
+                    for i, result in enumerate(results[:3]):
+                        print(f"   Result {i}: {result.get('url', 'unknown')[:50]}...")
+                else:
+                    print(f"❌ Crawl not successful: {data}")
+                    return False
+            else:
+                print(f"❌ REST API returned {response.status_code}")
+                print(f"   Response: {response.text[:500]}")
+                return False
+                
+        print("\n✅ REST API test completed successfully!")
+        return True
+        
+    except Exception as e:
+        print(f"❌ REST API test failed: {e}")
+        import traceback
+        traceback.print_exc()
+        return False
+
+
+async def main():
+    """Run all tests."""
+    print("\n🧪 TESTING COMPLETE FIX FOR DOCKER FILTER AND JSON ISSUES")
+    print("=" * 60)
+    print("Make sure the server is running with the updated code!")
+    print("=" * 60)
+    
+    results = []
+    
+    # Test 1: Docker client
+    docker_passed = await test_with_docker_client()
+    results.append(("Docker Client", docker_passed))
+    
+    # Test 2: REST API
+    rest_passed = await test_with_rest_api()
+    results.append(("REST API", rest_passed))
+    
+    # Summary
+    print("\n" + "=" * 60)
+    print("FINAL TEST SUMMARY")
+    print("=" * 60)
+    
+    all_passed = True
+    for test_name, passed in results:
+        status = "✅ PASSED" if passed else "❌ FAILED"
+        print(f"{test_name:20} {status}")
+        if not passed:
+            all_passed = False
+    
+    print("=" * 60)
+    if all_passed:
+        print("🎉 ALL TESTS PASSED! Both issues are fully resolved!")
+        print("\nThe fixes:")
+        print("1. Filter serialization: Fixed by not serializing private __slots__")
+        print("2. JSON serialization: Fixed by removing property descriptors from model_dump()")
+    else:
+        print("⚠️ Some tests failed. Please check the server logs for details.")
+    
+    return 0 if all_passed else 1
+
+
+if __name__ == "__main__":
+    import sys
+    sys.exit(asyncio.run(main()))

tests/docker/test_server_requests.py

@@ -635,7 +635,209 @@ class TestCrawlEndpoints:
             pytest.fail(f"LLM extracted content parsing or validation failed: {e}\nContent: {result['extracted_content']}")
         except Exception as e: # Catch any other unexpected error
             pytest.fail(f"An unexpected error occurred during LLM result processing: {e}\nContent: {result['extracted_content']}")
-            
+
+
+    # 7. Error Handling Tests
+    async def test_invalid_url_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for invalid URLs."""
+        payload = {
+            "urls": ["invalid-url", "https://nonexistent-domain-12345.com"],
+            "browser_config": {"type": "BrowserConfig", "params": {"headless": True}},
+            "crawler_config": {"type": "CrawlerRunConfig", "params": {"cache_mode": CacheMode.BYPASS.value}}
+        }
+        
+        response = await async_client.post("/crawl", json=payload)
+        # Should return 200 with failed results, not 500
+        print(f"Status code: {response.status_code}")
+        print(f"Response: {response.text}")
+        assert response.status_code == 500
+        data = response.json()
+        assert data["detail"].startswith("Crawl request failed:")
+
+    async def test_mixed_success_failure_urls(self, async_client: httpx.AsyncClient):
+        """Test handling of mixed success/failure URLs."""
+        payload = {
+            "urls": [
+                SIMPLE_HTML_URL,  # Should succeed
+                "https://nonexistent-domain-12345.com",  # Should fail
+                "https://invalid-url-with-special-chars-!@#$%^&*()",  # Should fail
+            ],
+            "browser_config": {"type": "BrowserConfig", "params": {"headless": True}},
+            "crawler_config": {
+                "type": "CrawlerRunConfig", 
+                "params": {
+                    "cache_mode": CacheMode.BYPASS.value,
+                    "markdown_generator": {
+                        "type": "DefaultMarkdownGenerator",
+                        "params": {
+                            "content_filter": {
+                                "type": "PruningContentFilter",
+                                "params": {"threshold": 0.5}
+                            }
+                        }
+                    }
+                }
+            }
+        }
+        
+        response = await async_client.post("/crawl", json=payload)
+        assert response.status_code == 200
+        data = response.json()
+        assert data["success"] is True
+        assert len(data["results"]) == 3
+        
+        success_count = 0
+        failure_count = 0
+        
+        for result in data["results"]:
+            if result["success"]:
+                success_count += 1
+            else:
+                failure_count += 1
+                assert "error_message" in result
+                assert len(result["error_message"]) > 0
+                
+        assert success_count >= 1  # At least one should succeed
+        assert failure_count >= 1  # At least one should fail
+
+    async def test_streaming_mixed_urls(self, async_client: httpx.AsyncClient):
+        """Test streaming with mixed success/failure URLs."""
+        payload = {
+            "urls": [
+                SIMPLE_HTML_URL,  # Should succeed
+                "https://nonexistent-domain-12345.com",  # Should fail
+            ],
+            "browser_config": {"type": "BrowserConfig", "params": {"headless": True}},
+            "crawler_config": {
+                "type": "CrawlerRunConfig", 
+                "params": {
+                    "stream": True,
+                    "cache_mode": CacheMode.BYPASS.value
+                }
+            }
+        }
+        
+        async with async_client.stream("POST", "/crawl/stream", json=payload) as response:
+            response.raise_for_status()
+            results = await process_streaming_response(response)
+        
+        assert len(results) == 2
+        
+        success_count = 0
+        failure_count = 0
+        
+        for result in results:
+            if result["success"]:
+                success_count += 1
+                assert result["url"] == SIMPLE_HTML_URL
+            else:
+                failure_count += 1
+                assert "error_message" in result
+                assert result["error_message"] is not None
+        
+        assert success_count == 1
+        assert failure_count == 1
+
+    async def test_markdown_endpoint_error_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for markdown endpoint."""
+        # Test invalid URL
+        invalid_payload = {"url": "invalid-url", "f": "fit"}
+        response = await async_client.post("/md", json=invalid_payload)
+        # Should return 400 for invalid URL format
+        assert response.status_code == 400
+        
+        # Test non-existent URL
+        nonexistent_payload = {"url": "https://nonexistent-domain-12345.com", "f": "fit"}
+        response = await async_client.post("/md", json=nonexistent_payload)
+        # Should return 500 for crawl failure
+        assert response.status_code == 500
+
+    async def test_html_endpoint_error_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for HTML endpoint."""
+        # Test invalid URL
+        invalid_payload = {"url": "invalid-url"}
+        response = await async_client.post("/html", json=invalid_payload)
+        # Should return 500 for crawl failure
+        assert response.status_code == 500
+
+    async def test_screenshot_endpoint_error_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for screenshot endpoint."""
+        # Test invalid URL
+        invalid_payload = {"url": "invalid-url"}
+        response = await async_client.post("/screenshot", json=invalid_payload)
+        # Should return 500 for crawl failure
+        assert response.status_code == 500
+
+    async def test_pdf_endpoint_error_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for PDF endpoint."""
+        # Test invalid URL
+        invalid_payload = {"url": "invalid-url"}
+        response = await async_client.post("/pdf", json=invalid_payload)
+        # Should return 500 for crawl failure
+        assert response.status_code == 500
+
+    async def test_execute_js_endpoint_error_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for execute_js endpoint."""
+        # Test invalid URL
+        invalid_payload = {"url": "invalid-url", "scripts": ["return document.title;"]}
+        response = await async_client.post("/execute_js", json=invalid_payload)
+        # Should return 500 for crawl failure
+        assert response.status_code == 500
+
+    async def test_llm_endpoint_error_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for LLM endpoint."""
+        # Test missing query parameter
+        response = await async_client.get("/llm/https://example.com")
+        assert response.status_code == 422  # FastAPI validation error, not 400
+        
+        # Test invalid URL
+        response = await async_client.get("/llm/invalid-url?q=test")
+        # Should return 500 for crawl failure
+        assert response.status_code == 500
+
+    async def test_ask_endpoint_error_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for ask endpoint."""
+        # Test invalid context_type
+        response = await async_client.get("/ask?context_type=invalid")
+        assert response.status_code == 422  # Validation error
+        
+        # Test invalid score_ratio
+        response = await async_client.get("/ask?score_ratio=2.0")  # > 1.0
+        assert response.status_code == 422  # Validation error
+        
+        # Test invalid max_results
+        response = await async_client.get("/ask?max_results=0")  # < 1
+        assert response.status_code == 422  # Validation error
+
+    async def test_config_dump_error_handling(self, async_client: httpx.AsyncClient):
+        """Test error handling for config dump endpoint."""
+        # Test invalid code
+        invalid_payload = {"code": "invalid_code"}
+        response = await async_client.post("/config/dump", json=invalid_payload)
+        assert response.status_code == 400
+        
+        # Test nested function calls (not allowed)
+        nested_payload = {"code": "CrawlerRunConfig(BrowserConfig())"}
+        response = await async_client.post("/config/dump", json=nested_payload)
+        assert response.status_code == 400
+
+    async def test_malformed_request_handling(self, async_client: httpx.AsyncClient):
+        """Test handling of malformed requests."""
+        # Test missing required fields
+        malformed_payload = {"urls": []}  # Missing browser_config and crawler_config
+        response = await async_client.post("/crawl", json=malformed_payload)
+        print(f"Response: {response.text}")
+        assert response.status_code == 422  # Validation error
+        
+        # Test empty URLs list
+        empty_urls_payload = {
+            "urls": [],
+            "browser_config": {"type": "BrowserConfig", "params": {}},
+            "crawler_config": {"type": "CrawlerRunConfig", "params": {}}
+        }
+        response = await async_client.post("/crawl", json=empty_urls_payload)
+        assert response.status_code == 422  # "At least one URL required"
+
 if __name__ == "__main__":
     # Define arguments for pytest programmatically
     # -v: verbose output

tests/test_preserve_https_for_internal_links.py

@@ -0,0 +1,175 @@
+#!/usr/bin/env python3
+"""
+Final test and demo for HTTPS preservation feature (Issue #1410)
+
+This demonstrates how the preserve_https_for_internal_links flag
+prevents HTTPS downgrade when servers redirect to HTTP.
+"""
+
+import sys
+import os
+from urllib.parse import urljoin, urlparse
+
+def demonstrate_issue():
+    """Show the problem: HTTPS -> HTTP redirect causes HTTP links"""
+    
+    print("=" * 60)
+    print("DEMONSTRATING THE ISSUE")
+    print("=" * 60)
+    
+    # Simulate what happens during crawling
+    original_url = "https://quotes.toscrape.com/tag/deep-thoughts"
+    redirected_url = "http://quotes.toscrape.com/tag/deep-thoughts/"  # Server redirects to HTTP
+    
+    # Extract a relative link
+    relative_link = "/author/Albert-Einstein"
+    
+    # Standard URL joining uses the redirected (HTTP) base
+    resolved_url = urljoin(redirected_url, relative_link)
+    
+    print(f"Original URL:    {original_url}")
+    print(f"Redirected to:   {redirected_url}")
+    print(f"Relative link:   {relative_link}")
+    print(f"Resolved link:   {resolved_url}")
+    print(f"\n❌ Problem: Link is now HTTP instead of HTTPS!")
+    
+    return resolved_url
+
+def demonstrate_solution():
+    """Show the solution: preserve HTTPS for internal links"""
+    
+    print("\n" + "=" * 60)
+    print("DEMONSTRATING THE SOLUTION")
+    print("=" * 60)
+    
+    # Our normalize_url with HTTPS preservation
+    def normalize_url_with_preservation(href, base_url, preserve_https=False, original_scheme=None):
+        """Normalize URL with optional HTTPS preservation"""
+        
+        # Standard resolution
+        full_url = urljoin(base_url, href.strip())
+        
+        # Preserve HTTPS if requested
+        if preserve_https and original_scheme == 'https':
+            parsed_full = urlparse(full_url)
+            parsed_base = urlparse(base_url)
+            
+            # Only for same-domain links
+            if parsed_full.scheme == 'http' and parsed_full.netloc == parsed_base.netloc:
+                full_url = full_url.replace('http://', 'https://', 1)
+                print(f"  → Preserved HTTPS for {parsed_full.netloc}")
+        
+        return full_url
+    
+    # Same scenario as before
+    original_url = "https://quotes.toscrape.com/tag/deep-thoughts"
+    redirected_url = "http://quotes.toscrape.com/tag/deep-thoughts/"
+    relative_link = "/author/Albert-Einstein"
+    
+    # Without preservation (current behavior)
+    resolved_without = normalize_url_with_preservation(
+        relative_link, redirected_url,
+        preserve_https=False, original_scheme='https'
+    )
+    
+    print(f"\nWithout preservation:")
+    print(f"  Result: {resolved_without}")
+    
+    # With preservation (new feature)
+    resolved_with = normalize_url_with_preservation(
+        relative_link, redirected_url,
+        preserve_https=True, original_scheme='https'
+    )
+    
+    print(f"\nWith preservation (preserve_https_for_internal_links=True):")
+    print(f"  Result: {resolved_with}")
+    print(f"\n✅ Solution: Internal link stays HTTPS!")
+    
+    return resolved_with
+
+def test_edge_cases():
+    """Test important edge cases"""
+    
+    print("\n" + "=" * 60)
+    print("EDGE CASES")
+    print("=" * 60)
+    
+    from urllib.parse import urljoin, urlparse
+    
+    def preserve_https(href, base_url, original_scheme):
+        """Helper to test preservation logic"""
+        full_url = urljoin(base_url, href)
+        
+        if original_scheme == 'https':
+            parsed_full = urlparse(full_url)
+            parsed_base = urlparse(base_url)
+            # Fixed: check for protocol-relative URLs
+            if (parsed_full.scheme == 'http' and 
+                parsed_full.netloc == parsed_base.netloc and
+                not href.strip().startswith('//')):
+                full_url = full_url.replace('http://', 'https://', 1)
+        
+        return full_url
+    
+    test_cases = [
+        # (description, href, base_url, original_scheme, should_be_https)
+        ("External link", "http://other.com/page", "http://example.com", "https", False),
+        ("Already HTTPS", "/page", "https://example.com", "https", True),
+        ("No original HTTPS", "/page", "http://example.com", "http", False),
+        ("Subdomain", "/page", "http://sub.example.com", "https", True),
+        ("Protocol-relative", "//example.com/page", "http://example.com", "https", False),
+    ]
+    
+    for desc, href, base_url, orig_scheme, should_be_https in test_cases:
+        result = preserve_https(href, base_url, orig_scheme)
+        is_https = result.startswith('https://')
+        status = "✅" if is_https == should_be_https else "❌"
+        
+        print(f"\n{status} {desc}:")
+        print(f"  Input: {href} + {base_url}")
+        print(f"  Result: {result}")
+        print(f"  Expected HTTPS: {should_be_https}, Got: {is_https}")
+
+def usage_example():
+    """Show how to use the feature in crawl4ai"""
+    
+    print("\n" + "=" * 60)
+    print("USAGE IN CRAWL4AI")
+    print("=" * 60)
+    
+    print("""
+To enable HTTPS preservation in your crawl4ai code:
+
+```python
+from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
+
+async with AsyncWebCrawler() as crawler:
+    config = CrawlerRunConfig(
+        preserve_https_for_internal_links=True  # Enable HTTPS preservation
+    )
+    
+    result = await crawler.arun(
+        url="https://example.com",
+        config=config
+    )
+    
+    # All internal links will maintain HTTPS even if 
+    # the server redirects to HTTP
+```
+
+This is especially useful for:
+- Sites that redirect HTTPS to HTTP but still support HTTPS
+- Security-conscious crawling where you want to stay on HTTPS
+- Avoiding mixed content issues in downstream processing
+""")
+
+if __name__ == "__main__":
+    # Run all demonstrations
+    demonstrate_issue()
+    demonstrate_solution() 
+    test_edge_cases()
+    usage_example()
+    
+    print("\n" + "=" * 60)
+    print("✅ All tests complete!")
+    print("=" * 60)