Linear Cache MCP

A cache-first MCP server for Linear that lets coding agents read from a local JSON cache instead of burning through API request budgets on every query.

What it does

• 13 MCP tools — search, read, create, update, and comment on Linear issues and projects
• Local JSON cache — fast planning reads with zero API cost
• Request budget tracking — prevents accidentally blowing through Linear's API rate limits
• Write-through safety — live-fetches entities before mutating, patches cache after writes
• Project-first model — issues live inside projects; projects hold durable context

Why cache-first?

Linear's API has rate limits. Coding agents that query Linear directly for every context read can exhaust them quickly. This server maintains a local mirror of your team's issues and projects, serves reads from cache when fresh, and only hits the Linear API for writes or when the cache is stale.

Quick start

npx linear-cache-mcp

Or for development:

cd packages/linear-cache-mcp
npm install
cp .env.example .env  # add your LINEAR_API_KEY
npm test
npm start

Full setup instructions are in the package README, including Pi coding agent integration.

Structure

Path	Purpose
packages/linear-cache-mcp/	MCP server
skills/linear-ops/	Agent skill template for Linear workflows
examples/pi-project/	Ready-to-copy configs for Pi projects
examples/generic-project/	Generic MCP client config template

Links

• npm: linear-cache-mcp
• Linear: linear.app
• Pi: pi-coding-agent

License

MIT