Linear MCP Server - Production Grade
Production-grade MCP server for Linear.app with enterprise resilience, performance optimization, and clean architecture.
Power your agent
การติดตั้งจากผู้เผยแพร่ ไม่ใช่เอนด์พอยต์ของ PlugMCP
URL แบบรีโมตหรือคำสั่งบนเครื่องของคุณ? การติดตั้งสองแบบต่างกันอย่างไร
รันบน PlugMCP
A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.
เปิดจากผู้เผยแพร่
Built on the foundation of cline/linear-mcp with 10 major enhancements and 5,116 lines of additional production code.
# Clone the repository
git clone https://github.com/locomotive-agency/linear-mcp.git
cd linear-mcp
# Install dependencies
npm install
# Build the server
npm run build
# Install globally
npm install -g @locomotive/linear-mcp
# Or install locally in your project
npm install @locomotive/linear-mcp
The key will look like: lin_api_xxxxxxxxxxxxxxxxxxxxxxxxxxxx
1. Locate your Claude Code config file:
~/.claude.json~/.claude.json%USERPROFILE%\.claude.json2. Add the MCP server configuration:
{
"mcpServers": {
"linear": {
"command": "node",
"args": ["/absolute/path/to/linear-mcp/build/index.js"],
"env": {
"LINEAR_API_KEY": "lin_api_your_key_here"
}
}
}
}
3. Restart Claude Code
4. Verify it's working:
Ask Claude: "List my Linear teams"
1. Locate your Claude Desktop config:
~/Library/Application Support/Claude/claude_desktop_config.json%APPDATA%\Claude\claude_desktop_config.json~/.config/Claude/claude_desktop_config.json2. Add MCP server configuration:
{
"mcpServers": {
"linear": {
"command": "node",
"args": ["/absolute/path/to/linear-mcp/build/index.js"],
"env": {
"LINEAR_API_KEY": "lin_api_your_key_here"
}
}
}
}
3. Restart Claude Desktop
4. Verify:
1. Open Cursor settings:
Cursor → Settings → MCPFile → Preferences → MCP2. Add server configuration:
Create/edit ~/.cursor/mcp.json (macOS/Linux) or %APPDATA%\.cursor\mcp.json (Windows):
{
"mcpServers": {
"linear": {
"command": "node",
"args": ["/absolute/path/to/linear-mcp/build/index.js"],
"env": {
"LINEAR_API_KEY": "lin_api_your_key_here"
}
}
}
}
3. Restart Cursor
4. Verify in Cursor chat:
Ask Cursor: "List my Linear teams using the MCP tools"
1. Locate Cline MCP settings:
~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json%APPDATA%\Code\User\globalStorage\saoudrizwan.claude-dev\settings\cline_mcp_settings.json~/.config/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json2. Add server configuration:
{
"mcpServers": {
"linear": {
"command": "node",
"args": ["/absolute/path/to/linear-mcp/build/index.js"],
"env": {
"LINEAR_API_KEY": "lin_api_your_key_here"
},
"disabled": false,
"autoApprove": []
}
}
}
3. Reload VS Code window
4. Verify:
Note: OpenAI doesn't directly support MCP servers, but you can:
Option 1: Use as Local API
Option 2: Use Claude Code as Proxy
Option 3: Direct Integration (requires development) Create an OpenAI plugin/action wrapper around this MCP server.
1. Locate Continue config:
~/.continue/config.json2. Add MCP server to tools section:
{
"tools": [
{
"type": "mcp",
"name": "linear",
"command": "node",
"args": ["/absolute/path/to/linear-mcp/build/index.js"],
"env": {
"LINEAR_API_KEY": "lin_api_your_key_here"
}
}
]
}
3. Restart Continue
1. Create/edit Windsurf MCP config:
~/.windsurf/mcp.json2. Add configuration:
{
"mcpServers": {
"linear": {
"command": "node",
"args": ["/absolute/path/to/linear-mcp/build/index.js"],
"env": {
"LINEAR_API_KEY": "lin_api_your_key_here"
}
}
}
}
3. Restart Windsurf
For any MCP-compatible client:
Configuration Template:
{
"mcpServers": {
"linear": {
"command": "node",
"args": ["/absolute/path/to/linear-mcp/build/index.js"],
"env": {
"LINEAR_API_KEY": "lin_api_your_key_here"
}
}
}
}
Requirements:
npm run build)build/index.js// Using Claude Code, Claude Desktop, or Cursor
"Create a new Linear issue in the LOCOMOTIVE team:
Title: Implement user authentication
Description: Add OAuth 2.0 authentication with Google and GitHub providers
Priority: High (2)
Estimate: 5 story points"
"Search for all 'In Progress' issues in the LOCOMOTIVE team and update them to 'In Review'"
"Create a new project called 'Q4 2025 Features' with these milestones:
- Alpha Release (Nov 1)
- Beta Release (Nov 15)
- Production (Dec 1)
Then create 5 initial issues for the project."
// MCP server automatically optimizes related queries
"Get the project details, all milestones, and all issues for project ABC-123"
// This uses query batching internally - 1 rate limit slot instead of 3
# Run all tests
npm test
# Run specific test suite
npm test -- rate-limiter.test.ts
# Run tests in watch mode
npm test:watch
# Run integration tests (requires LINEAR_API_KEY)
npm run test:integration
src/
├── auth/ # Modular authentication
│ ├── types.ts # Auth interfaces
│ ├── api-key-auth.ts # API Key implementation
│ ├── oauth-auth.ts # OAuth implementation
│ └── index.ts # Factory/adapter
├── core/
│ ├── handlers/ # Domain handlers with lifecycle hooks
│ ├── middleware/ # Rate limiting, retry logic
│ ├── errors/ # Domain-specific error types
│ └── types/ # TypeScript definitions
├── features/
│ ├── issues/ # Issue management
│ ├── projects/ # Project management
│ ├── teams/ # Team management
│ ├── milestones/ # Milestone management
│ ├── comments/ # Comment management
│ └── monitoring/ # Rate limit monitoring
└── graphql/
├── client.ts # GraphQL client with batching
├── queries.ts # Query definitions
└── mutations.ts # Mutation definitions
Tool: linear_get_rate_limit_status
"Check my Linear API quota status"
// Returns:
{
"warningLevel": "normal",
"usage": {
"requestsThisHour": 150,
"hourlyUsagePercent": 15.0
},
"quota": {
"remainingHour": 850,
"resetTime": "in 3245s"
}
}
linear_create_issueCreate a single issue with full field support (title, description, team, assignee, priority, estimate, project, custom display).
linear_create_issuesBatch operation - Create multiple issues in one API call. ~10x faster than individual creates.
linear_update_issueUpdate any issue field: title, description, assignee, priority, state, project, or milestone.
linear_bulk_update_issuesBatch operation - Update multiple issues with the same changes. Efficient with per-issue error handling.
linear_update_issue_milestoneAssign an issue to a milestone or remove milestone assignment (pass null/empty).
linear_search_issuesSearch issues with advanced filtering: query text, teams, assignees, states, priority. Supports pagination.
Tip: Keep first parameter ≤ 20 for best performance.
linear_delete_issueDelete a single issue by identifier (LOC-123) or UUID.
linear_delete_issuesBatch operation - Delete multiple issues at once.
linear_link_issuesCreate relationships between issues: "blocks", "related", or "duplicate".
linear_unlink_issuesRemove a relationship between issues.
linear_get_issue_commentsGet all comments for an issue, including threaded replies. Supports pagination and archived comments.
linear_create_commentAdd a comment to an issue or create threaded reply. Markdown supported. OAuth apps can set custom display name.
linear_create_project_with_issuesAtomic operation - Create a project and initial issues in one transaction.
Important: teamIds is an array, not a single teamId.
linear_get_projectGet detailed project information with rich text descriptions (documentContent support).
linear_search_projectsSearch for projects by exact name match.
linear_create_project_milestoneCreate a single milestone with name, description, target date, and sort order.
linear_create_project_milestonesBatch operation - Create multiple milestones for a project in one call.
linear_get_project_milestoneGet detailed milestone information including associated issues.
linear_get_project_milestonesList all milestones for a specific project with pagination.
linear_search_project_milestonesSearch milestones with filters: name, project, target date.
linear_update_project_milestoneUpdate milestone properties: name, description, target date, sort order, or move to different project.
linear_delete_project_milestoneDelete a milestone permanently.
linear_get_teamsList all teams with states, labels, and workflow details. Best Practice: Call once at session start and cache team IDs.
linear_get_userGet authenticated user information: ID, name, email, accessible teams. Use: Verify authentication, get user ID for assignments.
linear_get_rate_limit_statusReal-time monitoring of API quota and usage.
Returns:
Automatic Features:
linear_authInitialize OAuth flow (OAuth apps only). Most users should use API Key authentication instead.
linear_auth_callbackHandle OAuth authorization callback (OAuth apps only).
Priority Values:
0 = None1 = Urgent 🔴2 = High 🟠3 = Medium 🟡4 = Low 🔵Common State Names:
Estimate Points (story points):
"List my Linear teams"
→ Returns team names, keys, and UUIDs needed for other operations
"Create a Linear issue:
- Team: LOCOMOTIVE
- Title: Implement authentication system
- Description: Add OAuth 2.0 with Google and GitHub providers
- Priority: High (2)
- Estimate: 8 points
- Assignee: [email protected]
- Project: Phase 2 Features"
"Create 10 issues for the new feature:
1. Research requirements (estimate: 2)
2. Design architecture (estimate: 3)
3. Implement backend (estimate: 8)
..."
→ Uses linear_create_issues for efficient batch creation
"Find all urgent priority issues that are in Todo or In Progress state for the LOCOMOTIVE team"
→ Uses linear_search_issues with filters:
- teamIds: ["team-uuid"]
- states: ["Todo", "In Progress"]
- priority: 1
// Ask Claude Code, Cursor, or Claude Desktop:
"Create a new Linear project called 'Website Redesign' in the DESIGN team with these milestones:
1. Research & Planning (target: Dec 1, 2025)
2. Design Phase (target: Dec 15, 2025)
3. Development (target: Jan 15, 2026)
4. Launch (target: Feb 1, 2026)
Then create 3 initial issues:
- Conduct user research (milestone: Research & Planning)
- Create design mockups (milestone: Design Phase)
- Setup development environment (milestone: Development)"
"Find all issues with priority 'Urgent' in the 'Todo' state and move them to 'In Progress'"
// Uses:
// 1. linear_search_issues with filters
// 2. linear_bulk_update_issues for efficient batch update
"Create 3 related issues for the authentication feature:
1. Backend API (blocks the other two)
2. Frontend UI (blocked by #1)
3. Testing (related to both)
Link them appropriately."
// Uses:
// 1. linear_create_issues (batch)
// 2. linear_link_issues (create relationships)
"Check my Linear API quota and warn me if it's above 80%"
// Uses:
// linear_get_rate_limit_status
// Returns real-time quota with warning levels
Create a .env file (optional):
# Authentication (choose one)
LINEAR_API_KEY=lin_api_your_key_here
# Or OAuth (not fully implemented)
LINEAR_CLIENT_ID=your_client_id
LINEAR_CLIENT_SECRET=your_client_secret
LINEAR_REDIRECT_URI=http://localhost:3000/callback
# Optional: Rate limiting (defaults shown)
RATE_LIMIT_MAX_PER_HOUR=1000
RATE_LIMIT_MAX_PER_MINUTE=100
RATE_LIMIT_SAFETY_THRESHOLD=0.9 # Queue at 90%
# Optional: Retry configuration
RETRY_MAX_ATTEMPTS=5
RETRY_INITIAL_DELAY_MS=100
Configure via MCP settings:
{
"mcpServers": {
"linear": {
"command": "node",
"args": ["/path/to/linear-mcp/build/index.js"],
"env": {
"LINEAR_API_KEY": "your_key",
"RATE_LIMIT_SAFETY_THRESHOLD": "0.85",
"RETRY_MAX_ATTEMPTS": "3"
}
}
}
}
Automatically batches related queries for efficiency:
// When you ask:
"Get project details, milestones, and all issues for Project X"
// The server automatically batches these queries:
// - Uses 1 rate limit slot instead of 3
// - Executes with coordinated error handling
// - Returns all results together
Extensible handler system for cross-cutting concerns:
// Built-in hooks:
- Logging hook: Logs all operations automatically
- Metrics hook: Tracks duration and success rates
// Easy to add custom hooks:
- Caching hook: Cache frequently accessed data
- Audit hook: Track all operations for compliance
Type-safe error handling with structured logging:
try {
await linear.createIssue({ priority: 99 }); // Invalid
} catch (error) {
if (error instanceof ValidationError) {
console.log(`Field: ${error.field}`); // "priority"
console.log(`Value: ${error.value}`); // 99
console.log(`Valid: ${error.validValues}`); // [0,1,2,3,4]
}
}
// Structured logging:
{
"timestamp": "2025-10-29T...",
"level": "ERROR",
"type": "ValidationError",
"code": "VALIDATION_INVALID_VALUE",
"message": "Invalid priority value",
"context": {
"field": "priority",
"value": 99,
"validValues": [0,1,2,3,4]
}
}
Added Code: 5,116 lines beyond original
# Install dependencies
npm install
# Build
npm run build
# Run tests
npm test
# Watch mode
npm test:watch
# Type checking
npm run type-check
linear-mcp/
├── src/
│ ├── auth/ # Authentication (modular)
│ ├── core/ # Core systems
│ ├── features/ # Domain features
│ ├── graphql/ # GraphQL client
│ └── __tests__/ # Test suites
├── build/ # Compiled output
├── docs/ # Documentation
│ ├── IMPROVEMENTS.md
│ ├── PROJECT-COMPLETE.md
│ └── MIGRATION-PLAN.md
└── package.json
See CONTRIBUTING.md for:
Check build exists:
ls -lh /path/to/linear-mcp/build/index.js
Verify config path is absolute (not relative):
{
"args": ["/absolute/path/to/linear-mcp/build/index.js"] // ✅ Good
"args": ["./build/index.js"] // ❌ Won't work
}
Check Linear API key:
# Test API key manually
curl -H "Authorization: Bearer lin_api_your_key" \
https://api.linear.app/graphql \
-d '{"query": "{ viewer { name } }"}'
Symptom: "Authentication not initialized" or "Unauthorized"
Solutions:
Symptom: Console warnings about approaching limits
This is normal! The monitoring system warns you at:
Action: The rate limiter automatically queues requests, no action needed.
To monitor:
"Check my Linear API rate limit status"
# Clear cache and rebuild
rm -rf build/ node_modules/
npm install
npm run build
npm test
Verify server is running:
ls -lh build/index.js)See Linear API Documentation for field details.
Common Parameters:
teamId - Team UUID (required for most operations)priority - 0=None, 1=Urgent, 2=High, 3=Medium, 4=LowstateId - Workflow state UUIDestimate - Story points (typically 1, 2, 3, 5, 8)Get Team ID:
"List my Linear teams"
// Returns team IDs and names
This project builds upon the excellent foundation of cline/linear-mcp.
Original Authors: Cline team Enhanced By: LOCOMOTIVE Agency
See ATTRIBUTION.md for detailed credits.
MIT License - see LICENSE file for details.
See CONTRIBUTING.md for contribution guidelines.
✅ Rate limiting ensures you never hit API limits ✅ Automatic retry recovers from 90%+ of failures ✅ Real-time monitoring prevents surprises
✅ Query batching reduces API calls by 67% ✅ Optimized operations for complex workflows ✅ Efficient batch processing
✅ Modular architecture (easy to extend) ✅ Type-safe error handling ✅ Comprehensive tests (203 passing) ✅ Zero breaking changes (backward compatible)
✅ Lifecycle hooks for logging, metrics, caching ✅ Structured error logging ✅ Observable operations ✅ Production-ready
Version: 2.0.0 Status: Production Ready 🚀 Built with: ❤️ and ZERO MOCK compliance
Production-grade MCP server for Linear.app with enterprise resilience, performance optimization, and clean architecture.
เครื่องมือ
ยังไม่มีรายการเครื่องมือ เซิร์ฟเวอร์แบบรีโมตจะได้รายการหลังการตรวจสอบสถานะครั้งแรก
FAQ
คัดลอกคอนฟิก stdio สำหรับ Cursor, Claude Desktop, Claude Code หรือ VS Code ได้ในหน้านี้ คอนฟิกนี้ใช้การติดตั้งของผู้เผยแพร่ ไม่ใช่พร็อกซีของ PlugMCP
ระเบียนนี้ไม่ได้ระบุตัวแปรสภาพแวดล้อมที่จำเป็น
Linear ทำงานแบบโลคัลผ่าน stdio
ตรวจสอบล่าสุดเมื่อ 2026-10-28 Production-grade MCP server for Linear.app with enterprise resilience, performance optimization, and clean architecture.