MCP Index
Project ManagementFreeactiveReviewed by MCPIndex

Linear MCP Server

MCP tool connecting AI to Linear for task and project management using natural language.

View on GitHubView on NPMBrowse all tools

Looking for more MCP servers? Browse the full MCP tools directory or explore more tools in Project Management.

OverviewConfigSetupFAQRelated
Reviewed by MCPIndex

MCPIndex assessment

Linear MCP is the cleanest agent-to-work-queue integration available — Linear's API discipline (typed GraphQL, real cycle semantics) maps almost perfectly onto agent tool use. The two traps are grounding and search scope: ungrounded writes misfile silently, and default search hides archived issues, manufacturing duplicates. Mandate the search-before-create and ground-first prompting habits and this server will delete an entire category of standup overhead from your week.

Quick overview

An issue tracker is a work queue with a guilt-based UI, and Linear MCP removes the human from the critical path of queue hygiene. That is the real story here: not that Claude can create tickets — any webhook can do that — but that the agent can now traverse the full planning graph (cycles, projects, estimates, dependencies, labels) with GraphQL-precision and maintain it as a coherent system instead of a graveyard of stale issues. Connected via the hosted OAuth endpoint (mcp.linear.app), the agent triages incoming reports against existing issues to kill duplicates at birth, scopes bug reports into properly-estimated tickets with the right team and labels, and answers the question every standup fumbles: what actually moved this cycle. The architectural nuance worth respecting: Linear's API speaks in opaque internal IDs while humans speak in display keys (ENG-123), and the translation layer between those two vocabularies is where every broken Linear automation ever written goes to die. The MCP server handles that mapping natively — but only if you stop fighting it with raw IDs in your prompts.

Best for

The teams and workflows that benefit most from this tool.

Automated bug triage with duplicate detection

Cycle review and standup summary generation

Converting chat discussions into properly-scoped tickets

Cross-project dependency auditing

What this MCP server is best for

  • Reviewing Linear issues, cycles, team backlogs, and project status through Linear MCP Server without leaving your planning workflow.
  • Useful for checking blocked tickets, summarizing active work, finding ownership gaps, and updating issue context during sprint planning or standups.
  • Best for product and engineering teams that run execution in Linear and want the assistant to help with prioritization, triage, and work tracking.

When to choose it

Choose Linear MCP Server when you want an MCP server focused on Project Management and need tighter integration with your existing tools.

Good fit

linearproject-managementissuestasksagile

Limitations

Things to watch before choosing this tool.

Agent inherits the connecting user role — admin sessions grant admin powers

Default search excludes archived and completed issues (duplicate factory)

Display keys are not API IDs — ungrounded writes misfile silently

Linear-only; no value if your team lives in Jira

Linear MCP Server Configuration

Use the following configuration as a starting point for Claude Desktop or any compatible MCP client, then replace placeholder credentials with your own values.

claude_desktop_config.json
{
  "mcpServers": {
    "linear": {
      "url": "https://mcp.linear.app/mcp"
    }
  }
}

How to set up Linear MCP Server

These setup steps cover the typical installation flow for this MCP server.

  1. 1

    Add the remote server URL https://mcp.linear.app/mcp to your client — Linear operates the hosted endpoint; no local process or API key management required.

  2. 2

    Complete the OAuth flow and authorize with your Linear account — the agent inherits YOUR role permissions, so a workspace admin connecting it hands the agent admin-grade issue powers.

  3. 3

    Pro-Tip: The fatal flaw 90% of developers commit is referencing teams and issues by display names in prompts (the backend team, ENG-123 from memory) and trusting the agent to resolve them. Display keys and names are not API identifiers — when the agent guesses, it can file issues into the wrong team with total confidence and zero errors. Fix: run one grounding prompt at session start — list my teams and their current cycles — and let the agent cache real identifiers before any write operation. Every wrong-team ticket in existence was created by a skipped grounding step.

  4. 4

    Enable the duplicate-check habit: before any create, prompt the agent to search existing issues by keywords — Linear's fast search makes this nearly free and it is the difference between triage and spam.

  5. 5

    If the agent will only read (standup summaries, cycle reviews), create a dedicated Linear viewer account for it instead of using your admin session.

  6. 6

    Verify with: What are the open issues in my current cycle, grouped by assignee? A correct setup returns live cycle data in seconds.

Compatibility

Supported environments for Linear MCP Server.

Claude Desktop

Yes

Cursor

Yes

VS Code

Yes

Hosted OAuth endpoint by Linear. Ground the agent with a team-and-cycle lookup before any write operation; default search hides archived issues.

Frequently asked questions

Common questions for Linear MCP Server.

The agent insists an issue does not exist and creates a duplicate — while the issue is right there in Linear. No error ever appears. Why?

The silent failure of Linear automation, with two compounding causes. First: archived and completed issues are excluded from default list and search scopes, so the agent queries the live set, finds nothing, and concludes creation is safe. Second: display-key confusion — the human says ENG-123, the API wants an internal UUID, and a fuzzy resolution step silently maps to nothing. The engineering fix: before any create operation, require a search that explicitly includes archived and completed states (the server exposes filters for this), and reference issues by full identifier only after a grounding lookup. A duplicate ticket is never a create bug — it is always a search-scope bug.

Can the agent move issues through custom workflow states?

Yes — Linear's workflow states are per-team and fully API-addressable, including custom states. But state names vary per team, so the agent must look up the team's state list rather than assume a universal Done exists.

Does it respect cycle boundaries automatically?

Only if prompted to. Creating an issue does not auto-assign it to the active cycle — the agent must explicitly resolve the current cycle and set it. Ungrounded creates land in the backlog by default, which is either correct behavior or silent mis-filing depending on your intent.

Hosted endpoint vs community self-hosted servers?

The hosted mcp.linear.app endpoint is maintained by Linear, OAuth-secured, and tracks API changes automatically. Community GraphQL wrappers require personal API keys (which act as you, 24/7, from any machine holding them) and rot as the schema evolves. Use hosted unless you need to strip write tools at the transport layer.

Linear MCP Server vs Competitors

FeatureLinear MCP ServerCompetitor
Planning graph awareness✅ Native cycles, projects, estimates, and dependencies❌ Atlassian REST wrappers flatten sprints into raw ticket lists
Duplicate prevention✅ In-session search-before-create with archived-state filters❌ Email-to-ticket creates duplicates by design
Identifier handling✅ Resolves display keys to internal IDs natively❌ Raw API scripts force you to manage UUID mapping
Auth model✅ OAuth scoped to your session❌ Personal API tokens with standing full-account access
Enterprise process depth❌ Opinionated Linear workflow only✅ Jira's configurable workflows cover every enterprise pathology

Related Guides

How to Install MCP Servers

Cross-client installation guide for Claude Desktop, Cursor, and VS Code.

Claude Desktop MCP Setup

Beginner-friendly setup guide for Claude Desktop MCP.

Best Open Source MCP Tools on GitHub

Compare popular open source MCP tools by use case and setup difficulty.