Jira MCP Server
A clean, reliable Model Context Protocol server for Jira integration, enabling issue management, search, and more via natural language.
README
Jira MCP Server
A clean, reliable Model Context Protocol server for Jira integration, designed for use with Claude Desktop and other MCP clients.
Why This Exists
The existing Jira MCP servers have quality issues:
- Pydantic dependency conflicts
- Deprecated API usage (v2 instead of v3)
- Poor error handling
- Unreliable connections
This implementation focuses on:
- ✅ Modern Jira REST API v3
- ✅ Minimal, stable dependencies
- ✅ Proper error handling
- ✅ Multi-instance support
- ✅ Type safety throughout
Features
Core Operations
- 🔍 Search issues with JQL (Jira Query Language)
- 📄 Get issue details with full field information
- ➕ Create issues with descriptions, labels, parent/epic links, and custom fields
- ✏️ Update issues (summary, description, labels, etc.)
- 💬 Add comments to issues
- 🔄 Transition issues between statuses
- 📋 List projects accessible to the user
Advanced Features
- 🔗 Link issues with relationships (Relates, Blocks, Duplicates, etc.)
- 📊 Epic management - Create tasks under epics and query epic children
- 🔄 Get available transitions before transitioning issues
- 👥 User management - Search users and assign issues
- 🎯 Assignee control - Assign or unassign issues to team members
Multi-Instance Support
Connect to multiple Jira instances simultaneously:
- Different organizations
- Personal and work accounts
- Multiple teams
Quick Start
1. Installation
Option A: Install from PyPI (Recommended)
pip install jira-mcp-simple
Note: If jira-mcp command is not found, add Python's bin to PATH:
export PATH="$HOME/.local/bin:$PATH"
Option B: Install from source
git clone https://github.com/Positronic-AI/jira-mcp.git
cd jira-mcp
pip install -e .
2. Get Your Jira API Token
- Go to https://id.atlassian.com/manage-profile/security/api-tokens
- Click "Create API token"
- Give it a name (e.g., "Claude Desktop MCP")
- Copy the token
3. Test Connection
export JIRA_MYCOMPANY_URL="https://your-company.atlassian.net"
export JIRA_MYCOMPANY_EMAIL="your.email@company.com"
export JIRA_MYCOMPANY_TOKEN="your_api_token_here"
jira-mcp --test-connection mycompany
You should see:
✓ Connected successfully!
User: Your Name (your@email.com)
Account ID: 123abc...
Accessible projects: N
4. Configure Claude Desktop
Edit your Claude Desktop config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
Add this configuration:
{
"mcpServers": {
"jira": {
"command": "jira-mcp",
"args": ["--instance", "mycompany"],
"env": {
"JIRA_MYCOMPANY_URL": "https://your-company.atlassian.net",
"JIRA_MYCOMPANY_EMAIL": "your.email@company.com",
"JIRA_MYCOMPANY_TOKEN": "your_api_token_here"
}
}
}
}
Note: If you installed in a virtual environment, use the full path to jira-mcp, e.g., /path/to/venv/bin/jira-mcp
5. Restart Claude Desktop
Completely quit and reopen Claude Desktop.
6. Test It!
Try these commands in Claude:
List my Jira projects
Search for issues assigned to me
Show me PROJECT-123
Create a new task in PROJECT called "Test MCP integration"
Usage Examples
Natural Language Commands
Once configured, you can use natural language with Claude:
Basic Operations:
- "Find all open bugs in the MOBILE project"
- "Show me high priority issues assigned to me"
- "Create a task about fixing the login page"
- "Add a comment to PROJ-42 saying the fix is deployed"
- "Move PROJ-42 to In Progress"
- "Update the summary of PROJ-42 to include more details"
Advanced Operations:
- "Create a new epic called 'User Authentication System'"
- "Create a task under epic PROJ-10 for implementing login"
- "Link PROJ-42 to PROJ-50 with 'Blocks' relationship"
- "Show me all tasks under epic PROJ-10"
- "What transitions are available for PROJ-42?"
- "Search for user john.doe and assign PROJ-42 to them"
- "Unassign PROJ-42"
Direct JQL Queries
Search Jira with: project = PROJ AND status = "In Progress" ORDER BY priority DESC
Multi-Instance Configuration
To connect multiple Jira instances:
{
"mcpServers": {
"jira-work": {
"command": "jira-mcp",
"args": ["--instance", "work"],
"env": {
"JIRA_WORK_URL": "https://company.atlassian.net",
"JIRA_WORK_EMAIL": "you@company.com",
"JIRA_WORK_TOKEN": "token1"
}
},
"jira-personal": {
"command": "jira-mcp",
"args": ["--instance", "personal"],
"env": {
"JIRA_PERSONAL_URL": "https://personal.atlassian.net",
"JIRA_PERSONAL_EMAIL": "you@personal.com",
"JIRA_PERSONAL_TOKEN": "token2"
}
}
}
}
Claude will know which instance to use based on context.
Architecture
jira-mcp/
├── server.py # Main MCP server
├── jira_client.py # Jira REST API v3 wrapper
├── config.py # Configuration management
├── requirements.txt # Python dependencies
└── examples/ # Configuration examples
Key Design Decisions
Minimal Dependencies: Only essential, stable packages (mcp, httpx, pydantic, python-dotenv)
Modern API: Uses Jira REST API v3 (/rest/api/3/search/jql) instead of deprecated v2
Type Safety: Full type hints throughout for better IDE support and fewer bugs
Error Handling: Clear error messages propagated from Jira API
Multi-Instance: Environment-based configuration for multiple Jira instances
Available Tools
The server exposes these MCP tools:
Core Tools
| Tool | Description |
|---|---|
jira_search |
Search issues using JQL with pagination |
jira_get_issue |
Get detailed issue information |
jira_create_issue |
Create new issues with custom fields and parent links |
jira_update_issue |
Update existing issue fields |
jira_add_comment |
Add comments to issues |
jira_transition_issue |
Change issue status |
jira_list_projects |
List accessible projects |
Advanced Tools
| Tool | Description |
|---|---|
jira_link_issues |
Create relationships between issues (Relates, Blocks, etc.) |
jira_get_epic_issues |
Get all issues belonging to an epic |
jira_get_transitions |
Get available transitions for an issue |
jira_search_users |
Search users by name or email |
jira_assign_issue |
Assign or unassign issues to users |
Troubleshooting
Connection fails
- Verify your API token is correct
- Check the Jira URL has no trailing slash
- Ensure your email matches the Atlassian account
- Test with:
jira-mcp --test-connection <instance>
Claude Desktop doesn't see the server
- Verify JSON config is valid
- Use absolute paths (not
~/or relative) - Restart Claude Desktop completely
- Check Claude Desktop logs for errors
"Module not found" errors
- Ensure
jira-mcp-simpleis installed:pip install jira-mcp-simple - If using a virtual environment, ensure it's activated or use the full path
"Field cannot be set" errors
- Not all fields are available on all projects
- Check your project's issue type screen configuration
- Common issue:
priorityfield not on screen (make it optional)
Contributing
Contributions are welcome! See CONTRIBUTING.md for guidelines.
Areas we'd especially appreciate help:
- Unit and integration tests
- Jira Data Center support
- Attachment handling
- Sprint/board operations
- Custom field improvements
Development
# Clone and setup
git clone https://github.com/Positronic-AI/jira-mcp.git
cd jira-mcp
pip install -e ".[dev]"
# Test connection
export JIRA_TEST_URL="https://test.atlassian.net"
export JIRA_TEST_EMAIL="test@example.com"
export JIRA_TEST_TOKEN="test_token"
jira-mcp --test-connection test
License
MIT License - see LICENSE for details.
Acknowledgments
- Built for use with Claude Desktop
- Uses the Model Context Protocol
- Inspired by the need for reliable Jira automation
Support
Note: This is an independent project and is not affiliated with Atlassian or Anthropic.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。