nia-mcp-server 1.0.9__tar.gz → 1.0.11__tar.gz

{nia_mcp_server-1.0.9 → nia_mcp_server-1.0.11}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nia-mcp-server
-Version: 1.0.9
+Version: 1.0.11
 Summary: Nia Knowledge Agent
 Project-URL: Homepage, https://trynia.ai
 Project-URL: Documentation, https://docs.trynia.ai

{nia_mcp_server-1.0.9 → nia_mcp_server-1.0.11}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "nia-mcp-server"
-version = "1.0.9"
+version = "1.0.11"
 description = "Nia Knowledge Agent"
 readme = "README.md"
 requires-python = ">=3.8"

{nia_mcp_server-1.0.9 → nia_mcp_server-1.0.11}/src/nia_mcp_server/__init__.py

@@ -2,4 +2,4 @@
 NIA MCP Server - Proxy server for NIA Knowledge Agent
 """
 
-__version__ = "1.0.9"
+__version__ = "1.0.11"

{nia_mcp_server-1.0.9 → nia_mcp_server-1.0.11}/src/nia_mcp_server/api_client.py

@@ -402,7 +402,11 @@ class NIAApiClient:
         url_patterns: List[str] = None,
         exclude_patterns: List[str] = None,
         max_age: int = None,
-        only_main_content: bool = True
+        only_main_content: bool = True,
+        wait_for: int = None,
+        include_screenshot: bool = None,
+        check_llms_txt: bool = None,
+        llms_txt_strategy: str = None
     ) -> Dict[str, Any]:
         """Create a new documentation/web data source."""
         try:
@@ -415,10 +419,18 @@ class NIAApiClient:
             # Add optional parameters
             if max_age is not None:
                 payload["max_age"] = max_age
-            # Always use markdown for documentation scraping
-            payload["formats"] = ["markdown"]
+            # Don't hardcode formats - let backend defaults apply
+            # This allows screenshots to be captured by default
             if only_main_content is not None:
                 payload["only_main_content"] = only_main_content
+            if wait_for is not None:
+                payload["wait_for"] = wait_for
+            if include_screenshot is not None:
+                payload["include_screenshot"] = include_screenshot
+            if check_llms_txt is not None:
+                payload["check_llms_txt"] = check_llms_txt
+            if llms_txt_strategy is not None:
+                payload["llms_txt_strategy"] = llms_txt_strategy
             
             response = await self.client.post(
                 f"{self.base_url}/v2/data-sources",

{nia_mcp_server-1.0.9 → nia_mcp_server-1.0.11}/src/nia_mcp_server/server.py

@@ -506,7 +506,11 @@ async def index_documentation(
     url_patterns: Optional[List[str]] = None,
     exclude_patterns: Optional[List[str]] = None,
     max_age: Optional[int] = None,
-    only_main_content: Optional[bool] = True
+    only_main_content: Optional[bool] = True,
+    wait_for: Optional[int] = None,
+    include_screenshot: Optional[bool] = None,
+    check_llms_txt: Optional[bool] = True,
+    llms_txt_strategy: Optional[str] = "prefer"
 ) -> List[TextContent]:
     """
     Index documentation or website for intelligent search.
@@ -517,6 +521,13 @@ async def index_documentation(
         exclude_patterns: Optional list of URL patterns to exclude from crawling (e.g., ["/blog/*", "/changelog/*"])
         max_age: Maximum age of cached content in seconds (for fast scraping mode)
         only_main_content: Extract only main content (removes navigation, ads, etc.)
+        wait_for: Time to wait for page to load in milliseconds (defaults to backend setting)
+        include_screenshot: Whether to capture full page screenshots (defaults to backend setting)
+        check_llms_txt: Check for llms.txt file for curated documentation URLs (default: True)
+        llms_txt_strategy: How to use llms.txt if found:
+            - "prefer": Start with llms.txt URLs, then crawl additional pages if under limit
+            - "only": Only index URLs listed in llms.txt
+            - "ignore": Skip llms.txt check (traditional behavior)
         
     Returns:
         Status of the indexing operation
@@ -525,6 +536,7 @@ async def index_documentation(
         - When started indexing, prompt users to either use check_documentation_status tool or go to app.trynia.ai to check the status.
         - By default, crawls the entire domain (up to 10,000 pages)
         - Use exclude_patterns to filter out unwanted sections like blogs, changelogs, etc.
+        - Screenshots are captured by default to provide visual context
     """
     try:
         client = await ensure_api_client()
@@ -536,7 +548,11 @@ async def index_documentation(
             url_patterns=url_patterns,
             exclude_patterns=exclude_patterns,
             max_age=max_age,
-            only_main_content=only_main_content
+            only_main_content=only_main_content,
+            wait_for=wait_for,
+            include_screenshot=include_screenshot,
+            check_llms_txt=check_llms_txt,
+            llms_txt_strategy=llms_txt_strategy
         )
         
         source_id = result.get("id")