mcp-redmine
Enables AI coding agents to interact with Redmine projects, issues, time tracking, and members through natural language via the Model Context Protocol.
README
mcp-redmine
Connect coding agents to Redmine through the Model Context Protocol. Use a normal API key; the agent can search issues, update tickets, log time, and manage project basics without opening the web UI. Source: github.com/mejtyas/mcp-redmine.
Docker image: mejtyas/mcp-redmine:latest (run with docker run -i for stdio).
Tools (17)
| Read | Create / update |
|---|---|
redmine_get_users |
redmine_create_issue |
redmine_get_current_user |
redmine_edit_issue |
redmine_get_projects |
redmine_create_fixed_version |
redmine_get_issue |
redmine_add_member_to_project |
redmine_search_issues |
redmine_remove_member_from_project |
redmine_get_fixed_versions |
redmine_log_time |
redmine_get_project_members |
redmine_upload_attachment |
redmine_get_time_entries |
redmine_add_issue_relation |
redmine_execute_custom_request |
— |
Together these cover most day-to-day workflows: projects, issues (look up, search, create, edit), versions, members, time logging, attachments, and relations. For typical agent tasks—triage, updates, notes, light reporting—that is roughly 95% of what you need without custom glue code.
Before you start
- In Redmine, open My account and create an API access key if you do not have one.
- Set credentials using the canonical names below (shell, Cursor
envblock, or inline Docker-eVAR=valueinargs—see Cursor). The server also acceptsredmine_urlandredmine_api_tokenas aliases.
| Variable | Value |
|---|---|
REDMINE_URL |
Base URL, e.g. https://redmine.example.com (no trailing slash required) |
REDMINE_API_KEY |
Your API key string |
Everything the agent does runs as the Redmine user tied to that key.
Run the server (Docker)
docker run --rm -i \
-e REDMINE_URL=https://your-redmine.example.com \
-e REDMINE_API_KEY=your_key \
mejtyas/mcp-redmine:latest
Company HTTPS deployment (Cursor url)
For a shared host (for example https://mcp-redmine.example.com/mcp), the server can keep one company Redmine URL and require a shared gate token plus each developer’s Redmine API key on every HTTP request.
- Redmine URL — either set
REDMINE_URLin the container environment, or setCOMPANY_REDMINE_URLinmcp_redmine/config.py(used whenREDMINE_URLis unset). - Server environment (deployment / GitLab CI variables, not in
mcp.json):MCP_TRANSPORT—streamable-http(recommended) orsseMCP_AUTH_TOKEN— long random secret shared inside the company (required for HTTP transport)- Optional:
MCP_HOST,MCP_PORT(default0.0.0.0/8000),MCP_PATH(default/mcp)
- Cursor — remote entry with two headers: company gate + personal Redmine key:
{
"mcpServers": {
"mcp-redmine": {
"url": "https://mcp-redmine.example.com/mcp",
"headers": {
"Authorization": "Bearer ${env:MCP_AUTH_TOKEN}",
"X-Redmine-API-Key": "${env:REDMINE_API_KEY}"
}
}
}
}
Use the same header names if you terminate TLS in front of the app; the MCP process must still receive Authorization: Bearer … and X-Redmine-API-Key: …. Do not commit real tokens.
Cursor
Open Settings → MCP, or edit ~/.cursor/mcp.json. Use a top-level mcpServers object.
Option A — env block (recommended)
Cursor sets env on the MCP server process. Use bare -e VAR in Docker args so those variables are forwarded into the container (no secrets duplicated in the run argument list).
{
"mcpServers": {
"mcp-redmine": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"REDMINE_URL",
"-e",
"REDMINE_API_KEY",
"mejtyas/mcp-redmine:latest"
],
"env": {
"REDMINE_URL": "https://your-redmine.example.com",
"REDMINE_API_KEY": "your_key"
}
}
}
}
Option B — inline -e in args
If you prefer everything in args, this shape also works:
{
"mcpServers": {
"mcp-redmine": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"REDMINE_URL=https://your-redmine.example.com/",
"-e",
"REDMINE_API_KEY=your_key",
"mejtyas/mcp-redmine:latest"
]
}
}
}
Replace placeholders with your values. Do not commit real URLs or keys.
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。