mcpField
U
uptodoc

uptodoc

Local
github.com/jlg-formationAI & Agents1Updated 79w ago
Repository

About uptodoc

uptodoc is a local MCP server that lets IDE-integrated AI agents query a configurable documentation database. It is intended for coding assistants that need recent or project-specific library and framework documentation.

Key features

  1. 01Query a documentation database
  2. 02Configure the documentation endpoint
  3. 03Integrate with IDE-based AI assistants

Use cases

  1. 01Retrieve recent library documentation
  2. 02Add project-specific knowledge to coding suggestions

MCP support

It implements MCP as a local server and provides documentation queries to compatible MCP clients.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "uptodoc": {
      "command": "npx",
      "args": [
        "-y",
        "@jlgformation/uptodoc"
      ],
      "env": {
        "UPTODOC_ENDPOINT": "<UPTODOC_ENDPOINT>"
      }
    }
  }
}
Run commandnpx -y @jlgformation/uptodoc
Environment variables to set
UPTODOC_ENDPOINT

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@jlgformation/uptodoc

Tools

1

Saved tool list · updated 3w ago

ParameterTypeRequiredDescription
librarystringYesLibrary name.

Details

From the project README

uptodoc

A lightweight MCP (Model Context Protocol) server for IDE-integrated AI agents (e.g., GitHub Copilot, Roocode, Cursor, Windsurf).

Overview

uptodoc provides a local server implementing the Model Context Protocol (MCP).
It allows AI assistants integrated into IDEs like VSCode to query a custom documentation database, offering more relevant and up-to-date information about libraries and frameworks.

Once configured in your IDE, the assistant will automatically communicate with uptodoc to enhance coding suggestions.

Typical use cases:

  • Fetching more recent or project-specific documentation
  • Improving AI suggestions in coding assistants
  • Extending Copilot, Roocode, Cursor, or Windsurf with custom knowledge

Requirements

  • Node.js version 22 or higher
  • VSCode version 1.99.3 or higher (or any compatible MCP client)

Make sure these tools are installed and up to date before using uptodoc.

Installation

You don't need to install uptodoc globally.
It is automatically launched by your IDE when configured correctly.

Make sure npx is available on your system (comes with Node.js).

Note: The package must be available on npm under the name @jlgformation/uptodoc.

IDE Configuration

To use uptodoc in VSCode, add the following to your settings (e.g., .vscode/settings.json) :

{
  // rest of the user settings.json config.
  "mcp": {
    "servers": {
      "uptodoc": {
        "type": "stdio",
        "command": "npx",
        "args": ["-y", "@jlgformation/uptodoc", "start"],
        "env": {
          "UPTODOC_ENDPOINT": "https://<yoursite.org>/<path>"
        }
      }
    }
  }
}

Put the Github Copilot AI Assistant in agent mode. Then start the MCP server on vscode.

You should be able to talk to the AI assistant. Do not hesitate to add in your prompt the magical formula: use uptodoc.

Example:

I want to install the tototructruc library. How to do ? use uptodoc.

As you can see above, the endpoint for the documentation can be configured. By default it is:

UPTODOC_ENDPOINT=https://raw.githubusercontent.com/jlg-formation/uptodoc/master/docs

And there are some stupid docs: https://github.com/jlg-formation/uptodoc/tree/master/docs

References

Author

Jean-Louis GUENEGO [email protected]

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.