MCP Servers
Last updated
Was this helpful?
Last updated
Was this helpful?
An MCP server, short for Model Context Protocol server, is a lightweight program that exposes specific capabilities to AI models via a standardized protocol, enabling AI to interact with external data sources and tools.
MCP servers allows you to extend BoltAI's capabilities with custom tools and commands.
Starting from v1.34, you can add any MCP server to BoltAI. Read the guide below to get started.
BoltAI follows the standard MCP configuration convention and stores all data in an mcp.json file. There are 2 ways to add a new MCP server to BoltAI.
Import from another app.
Edit the configuration file directly.
If you're already using an MCP client like Cursor or Claude desktop app, you can import the configuration in 2 simple steps.
Go to Settings > Plugins.
Click the elipsis button, select "Import from Cursor" or "Import from Claude".
Your MCP servers should show up on the plugin list.
Go to Settings > Plugins. Click the plus button. Update the file content and click Save Changes.
In BoltAI, an MCP server functions similarly to an AI plugin: simply enable it in the Plugin Dropdown, and you can start interacting with the server.
Here are a few useful resources about MCP to help you get started:
Is MCP configuration sync'ed between devices? No. MCP server configuration might contain sensitive data such as API key or password. Please copy, or symlink the configuration file if needed.
Where does BoltAI store the MCP configuration?
BoltAI stores all configuration in /Users/<user>/.boltai/mcp.json
Can I use a remote MCP server?
Yes, you can use mcp-remote to convert a remote MCP server to a local Stdio server.
