Outer Web MCP Server

About

What is this?

This is a Model Context Protocol (MCP) server that provides semantic search and exploration capabilities for the Outer Web blog collection. MCP allows LLM agents to access tools and data sources through a standardized protocol.

How do I use it?

This MCP server can be used with any MCP-compatible client. To connect:

HTTP Endpoint: https://outerweb.org/mcp/
Protocol: MCP over HTTP

Claude Code / Claude Desktop: https://docs.claude.com/en/docs/claude-code/mcp

Available Tools

search_posts

Search for blog posts semantically related to a query. Uses both text matching and semantic embeddings for high-quality results.

Parameters:

query (string, required) - The search query text
count (integer, optional) - Number of results to return (default: 10)
emphasis (string, optional) - Keywords to emphasize in search (comma-separated)

list_categories

List all available blog post categories in the Outer Web collection.

Parameters: None

explore_category

Explore posts within a specific category and discover related content.

Parameters:

category (string, required) - Category name (use list_categories to see options)
max_results (integer, optional) - Maximum number of posts to return (default: 5)

get_post_details

Get detailed information about a specific blog post by its URL.

Parameters:

url (string, required) - The full URL of the blog post

explore_similar

Find posts similar to a given post by its MD5 ID by exploring its neighbors in the link graph and semantic space.

Parameters:

post_id (string, required) - The MD5 ID of the post to find similar content for
count (integer, optional) - Number of similar posts to return (default: 5)

Note: To get the MD5 ID of a post, use get_post_details or extract it from search_posts results.

Example Usage

Search for Rust posts

await client.call_tool(
          "search_posts",
          arguments={
              "query": "rust embedded systems ESP32",
              "count": 15,
              "emphasis": "rust, embedded, microcontroller"
          }
      )

Explore a category

await client.call_tool(
          "explore_category",
          arguments={
              "category": "programming",
              "max_results": 10
          }
      )

Find similar posts

await client.call_tool(
            "explore_similar",
            arguments={
                "post_id": "a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6",
                "count": 5
            }
        )
            