Local MCP Server

Run the Obris MCP server on your machine using an API key. This connects your AI tools to your curated knowledge over stdio without going through the hosted server.

Don't want to read the docs? Talk to our Expert to get started.

This is an alternative to the remote MCP connectors which use OAuth and connect directly to your expert's remote MCP URL. The local server is a good option if you want to use an expert on your own machine. Use remote MCP if your team needs access or if you want the expert available across devices (including mobile and tablets).

ChatGPT does not support local stdio MCP servers. Use the remote MCP connector or a Custom GPT instead.

Connect your AI platform

Select your AI platform below to see the setup instructions.

Connect via the Claude Desktop app config.

In Claude Desktop, go to Settings > Developer and click Edit Config to open claude_desktop_config.json
Run which uvx in your terminal to get the full path to uvx

Add the Obris server to your config:

{
  "mcpServers": {
    "obris": {
      "command": "/full/path/to/uvx",
      "args": ["obris-mcp"],
      "env": {
        "OBRIS_API_KEY": "$OBRIS_API_KEY"
      }
    }
  }
}

Need an API key? Go to Experts, select your expert, and generate one from the API Keys tab.

Restart Claude Desktop

Set your API key

Add your API key to your shell profile so it is available in every terminal session. Add the following to your ~/.zshrc or ~/.bash_profile, then run source ~/.zshrc:

export OBRIS_API_KEY=your_api_key_here

Check the connection

Once connected, verify your expert is working by asking it a question. Try a prompt like:

What topics do I have in Obris?

Your expert should be able to list topics and the knowledge in them.

AI tools decide when to invoke MCP tools. If your expert does not pull from your Obris data, try mentioning Obris explicitly, for example "Using Obris, find my knowledge about..."