MCP ServerSSE

Linear MCP Server

Linear hosts an official MCP server for issues, projects, teams and comments. It speaks streamable HTTP at mcp.linear.app/mcp and SSE at mcp.linear.app/sse. Authenticate with a Linear personal API key from Settings → API, paste the URL and token into any MCP client, and your backlog becomes a live agent surface.

Try Linear in Agent Studio Linear MCP docs

Hosted URL

https://mcp.linear.app/sse

Suggested model

Claude Sonnet 4.5

Chat with 60+ AI models on the same workflow — switch to a different model mid-conversation and re-run the same prompt, or use Compare mode to put several side-by-side and balance quality vs. cost.

Auth

Linear personal API key (Settings → API → Personal API keys). Scopes should match the teams and actions you need.

What the Linear MCP server does

How models use it and what it is built for.

The Linear MCP server wraps Linear’s product-management API: listing teams, cycles, issues and projects; fetching a single issue or document; creating or updating work (using the vendor’s current save/create tools); and posting comments. Models can answer “what is blocked in the current cycle?”, “summarise what shipped last week”, or “create a P1 bug for the checkout failure in ENG”. Because Linear runs the server, you do not self-host a bridge — you only need API credentials with the right scopes. Agent Studio and other clients connect over HTTP or SSE using the same endpoint family.

Tools the Linear MCP server exposes

Typical tools an AI model can call. Exact names vary by version.

list_issues / get_issue — search and read issues, including filters and status
save_issue — create or update an issue (Linear’s unified write tool)
list_teams / list_cycles — discover teams, cycles, and workflow states
list_comments / create_comment — read or add discussion on issues
list_projects / get_project — navigate projects and roadmaps
list_issue_statuses / list_issue_labels — see workflow columns and tags

Example prompts to try

Copy any of these into MCP Agent Studio after connecting.

What issues are in progress in the current cycle for my team?
List unassigned P1 bugs in the backlog and suggest who might own them.
Summarise everything my team closed in the last 7 days with links to issues.
Create a new issue in ENG titled "Investigate 500s on /checkout" and set it to High priority.

Models on MCP Playground

This is not a single-model product: you get the same MCP connection with 60+ models (Claude, GPT, Gemini, DeepSeek, open-weight, and more), you can switch mid-conversation, and you can open Compare mode to run the same prompt against multiple models at once. The card above is a suggested starting point for this server — not the only choice.

Default pick for Linear

Claude Sonnet 4.5

Claude Sonnet 4.5 handles multi-step backlog reasoning and state changes reliably. The Linear template defaults to Haiku 4.5 for lighter triage; compare both in Agent Studio.

Open Compare mode (side-by-side)Try this server in Agent Studio

Check an AI agent can actually use the Linear MCP server

Listing tools proves the server is reachable, not that a model can work with it. Evals go further: they read every tool on the server, write a test suite from its real schemas, and run it — code decides pass/fail on the responses (schema conformance, error codes, pagination, result caps) while a scoring model grades plain-English tasks driven through the tools.

Get a pass/fail report per tool with the evidence behind each verdict — and replay the same suite after every schema change. Destructive tools are excluded from the run.

Run evals

Try the Linear MCP server in your browser

Open MCP Agent Studio with the connection pre-filled. Add your token, pick any of 60+ models, and start chatting — no install required.

Try Linear in Agent Studio

Linear MCP server — FAQ

Common questions about connecting, scoping and using it safely.

What is the Linear MCP server?

It is the official remote MCP service Linear hosts for its product. It exposes your workspace’s issues, teams, cycles, projects and comments as tools so an AI can read the backlog, update work and add comments the same way the Linear API does.

How do I authenticate the Linear MCP server?

Create a personal API key under Linear → Settings → API, with permissions for the teams and objects you want the agent to touch. In MCP clients that do not use OAuth, pass that key as a bearer token when connecting to the hosted URL.

Can the agent change or delete production issues?

Yes, if the API key can write issues in those teams. Start with a key scoped to a test team or a single project, review proposed issue updates before you approve runs in production, and use read-only or narrower keys until you trust the workflow.

Does the Linear MCP server work with Claude, ChatGPT, Cursor, or the browser?

Yes. Any MCP client that can reach the hosted URL with your token works — for example Claude Desktop, Cursor, or MCP Agent Studio in the browser. Use HTTP at mcp.linear.app/mcp or SSE at mcp.linear.app/sse depending on what your client supports.

Which model is best for Linear and backlog work?

Use Claude Sonnet 4.5 for planning, summarisation, and writing issue descriptions. Claude Haiku 4.5 is enough for short triage. Run two models on the same thread in Agent Studio to compare answers before you standardise.