Connect MCP Servers to Linear Agent
Linear Agent can connect to MCP servers to access tools and data outside of Linear.
This lets the agent bring external context into your workflows to investigate issues, plan projects, write specs, and draft updates grounded in your full context.
MCP works across Linear Agent chat, comments, and automations, and is enabled at the workspace level by Admins.
Current available servers that can be connected to:
- Amplitude
- Attio
- Better Stack
- Datadog
- GitHub
- Glean
- Granola
- HubSpot
- incident.io
- Intercom
- Jam
- Notion
- PostHog
- Sentry
- Slack
- Stripe
An admin must enable Agent MCP in Workspace Settings to allow users connect:
From the Workspace Settings an Admin can:
- Enable or disable MCP support for the workspace
- Restrict usage to approved MCP servers
- Manage workspace-connected MCP servers
To Enable servers on the Workspace:
- Open workspace settings > AI & Agents settings > Linear Agent
- Beside Enable MCP servers toggle it to “On”
- Click on Manage next to “Allowed MCP servers”
- Scroll down to MCP servers and select any that users and automations can connect to via Linear Agent
Select available servers for the workspace including any custom servers using a URL
To Enable a Custom Server on the Workspace
- Click + Custom URL…
- Select Custom URL…
- Enter your server URL (must start with
http://orhttps://) and complete the connection flow:
- No authentication → connects immediately
- OAuth supported → sign in and authorize
Once connected, the customer server will be available for users to connect to through their settings
How a user connects Linear Agent to an available MCP Server:
- Open Settings > Agent personalization
- Under MCP servers click on the “+” (Add server) button
- Choose a server and start the connection flow
- Complete authentication with the provider
- Return to Linear and confirm the server shows as connected
After a server is connected, users can reference it in natural language when using Linear Agent. The agent can use the connected MCP server to gather information or perform supported actions. For example:
Example Prompts
- Investigate this issue using Sentry and summarize the likely cause
- Check Datadog for recent errors related to this workflow
- Look up the latest notes about this customer in Intercom
- Search Notion for relevant documents and generate a guides ahead of launch
- Use GitHub to find the pull request that likely introduced this change