Oi Wiki

This MCP server augments your large language model with a rich OA-IO Wiki-style knowledge base by summarizing the 462 OI-Wiki pages and serving the most relevant wiki markdown through a semantic vector store. It enables your model to fetch concise, contextually relevant Wiki content during conversations.

How to use

You use this MCP server with your MCP client to query OI-Wiki-derived content by leveraging a local or remote vector store. The server runs a Python-based backend that retrieves the closest wiki embeddings and returns the corresponding markdown for your chat or application to render.

How to install

Prerequisites you need before installing this MCP server:

• You must have the MCP client runtime uv available on your machine.

Follow these concrete steps to set up and run the server locally:

Install and run the oi-wiki MCP server

{
  "mcpServers": {
    "oi_wiki": {
      "type": "stdio",
      "name": "oi_wiki",
      "command": "uv",
      "args": [
        "--directory",
        "<path of MCP servers>/mcp-oi-wiki",
        "run",
        "python",
        "main.py"
      ]
    }
  }
}

Clone the project and prepare the environment

1. Open a terminal and navigate to your MCP servers directory.

2. Clone the MCP server repository with submodules.

3. Ensure uv is installed and available in your PATH.

Configure the MCP client to use the oi_wiki server

{
  "mcpServers": {
    "oi_wiki": {
      "command": "uv",
      "args": [
        "--directory",
        "<path of MCP servers>/mcp-oi-wiki",
        "run",
        "python",
        "main.py"
      ]
    }
  }
}

Alternate update flow to refresh the knowledge base

This setup supports updating the local knowledge base by generating a new database and refreshing embeddings.

Notes on updates and data sources

• You can generate your own database file at db/oi-wiki.db by running the update scripts that fetch, summarize, and embed OI-Wiki pages.

• Place your Silicon flow API key into an api.key file to enable API-based processing during the update flow.

• Execute the update steps in sequence: request summaries, produce result.jsonl, then run the generation script to create a new db/oi-wiki.db.

Security and maintenance

Keep the API key and access controls secure. Regularly regenerate embeddings if the underlying wiki content changes significantly, and monitor the vector store size to maintain performance.