Grokipedia MCP Server

Instructions

Implementation Reference

• grokipedia_mcp/server.py:131-178 (handler)

  The implementation of the get_page tool, which fetches a page from Grokipedia.

  async def get_page(
      slug: Annotated[str, Field(description="Unique slug identifier of the page to retrieve")],
      max_content_length: Annotated[int, Field(description="Maximum length of content to return (default: 5000)", ge=100)] = 5000,
      ctx: Context[ServerSession, AppContext] | None = None,
  ) -> CallToolResult:
      """Get complete Grokipedia page with metadata, content preview, and citations.
  
      Use for: reading articles, getting overviews, checking citations and sources.
      Returns: title, description, content preview (truncated), citations list.
      Tips: Use get_page_content for full untruncated content. Slug comes from search results.
      """
      if ctx is None:
          raise ValueError("Context is required")
  
      await ctx.debug(f"Fetching page: '{slug}'")
  
      try:
          client = ctx.request_context.lifespan_context.client
          result = await client.get_page(slug=slug, include_content=True)
  
          if not result.found or result.page is None:
              await ctx.warning(f"Page not found: '{slug}', searching for alternatives")
              search_result = await client.search(query=slug, limit=5)
              if search_result.results:
                  suggestions = [f"{r.title} ({r.slug})" for r in search_result.results[:3]]
                  await ctx.info(f"Found {len(search_result.results)} similar pages")
                  raise ValueError(
                      f"Page not found: {slug}. Did you mean one of these? {', '.join(suggestions)}"
                  )
              raise ValueError(f"Page not found: {slug}")
  
          await ctx.info(f"Retrieved page: '{result.page.title}' ({slug})")
          
          page = result.page
          content_len = len(page.content) if page.content else 0
          is_truncated = content_len > max_content_length
          
          text_parts = [
              f"# {page.title}",
              "",
              f"**Slug:** {page.slug}",
          ]
          
          if page.description:
              text_parts.extend(["", f"**Description:** {page.description}", ""])
          
          if page.content:
              preview_length = min(1000, max_content_length)

• grokipedia_mcp/server.py:124-130 (registration)

  The registration of the get_page tool using the @mcp.tool decorator.

  @mcp.tool(
      annotations=ToolAnnotations(
          readOnlyHint=True,
          destructiveHint=False,
          idempotentHint=True
      )
  )