odoo_mcp
Connects AI agents in Cursor to an Odoo instance, enabling natural language management of projects, epics, tasks, and milestones with tags and priorities.
README
Odoo MCP Server
An MCP (Model Context Protocol) server that connects AI agents in Cursor to your Odoo instance. Tell your AI to "turn this PRD into tasks in Odoo" and it will create a full hierarchy of epics, tasks, and subtasks with tags, priorities, and assignments.
No cloning required — just configure and go.
Quick Start
1. Install uv
uv is a fast Python package manager that includes uvx for running tools directly from git repos.
curl -LsSf https://astral.sh/uv/install.sh | sh
2. Get Your Odoo API Key
- Log into your Odoo instance
- Go to My Profile (click your avatar → My Profile)
- Navigate to Preferences → Account Security
- Click New API Key, name it "Cursor MCP", and copy the key
3. Configure Cursor
Create or edit ~/.cursor/mcp.json:
{
"mcpServers": {
"odoo": {
"command": "/Users/YOUR_USERNAME/.local/bin/uvx",
"args": [
"--from",
"git+https://github.com/willy5xx/odoo_mcp.git",
"odoo-mcp"
],
"env": {
"ODOO_URL": "https://your-instance.odoo.com",
"ODOO_DB": "your-database-name",
"ODOO_USERNAME": "your-email@example.com",
"ODOO_API_KEY": "your-api-key"
}
}
}
}
Replace the placeholder values:
YOUR_USERNAME— your macOS username (runwhoamito check)ODOO_URL— your Odoo instance URL (e.g.,https://mycompany.odoo.com)ODOO_DB— your database name (often the subdomain, e.g.,mycompany)ODOO_USERNAME— your Odoo login emailODOO_API_KEY— the API key you created in step 2
4. Restart Cursor
Quit and reopen Cursor for the MCP server to load.
6. Test It
In Cursor, ask your AI:
- "Test the Odoo connection"
- "List my Odoo projects"
If it works, you're all set!
Important Notes
| Issue | Solution |
|---|---|
| uvx not found | Cursor can't find uvx via PATH. Always use the full absolute path: /Users/YOUR_USERNAME/.local/bin/uvx |
| SSL certificate errors | Usually handled automatically. If you're behind a corporate proxy, see Troubleshooting |
| ODOO_URL format | Use the base URL only (e.g., https://mycompany.odoo.com). Do NOT include /odoo or other paths |
| Changes not taking effect | Restart Cursor after editing ~/.cursor/mcp.json |
Usage Examples
Once configured, you can use natural language to manage Odoo:
"List my Odoo projects"
"Read @specs/feature-prd.md and create tasks in the 'Mobile App' project.
Break it into epics with subtasks, tag by feature area, and set priorities."
"Show me tasks in the Backend project and update the auth task to urgent priority"
"Create a new task in the API project: implement rate limiting for the /users endpoint"
"Add a subtask to task #123: write unit tests for the validation logic"
Available Tools
| Tool | Description |
|---|---|
odoo_test_connection |
Verify credentials and connectivity |
odoo_list_projects |
List all projects with metadata |
odoo_create_project |
Create a new project |
odoo_update_project |
Update project fields (name, manager, active status) |
odoo_get_project_stages |
Get kanban stages for a project |
odoo_create_task |
Create a single task (supports subtasks via parent_id) |
odoo_create_tasks_batch |
Create multiple tasks at once |
odoo_list_tasks |
List tasks in a project |
odoo_update_task |
Update any field on an existing task |
odoo_post_task_message |
Post progress updates to task chatter |
odoo_search_users |
Find users by name/email for assignment |
odoo_list_tags |
List existing project tags |
odoo_create_milestone |
Create project milestones |
odoo_list_milestones |
List milestones in a project |
odoo_search_records |
Generic search on any Odoo model |
odoo_create_record |
Generic create on any Odoo model |
odoo_update_record |
Generic update for any Odoo model |
odoo_delete_record |
Generic delete for any Odoo model |
How PRD → Tasks Works
When you ask your AI to turn a PRD into Odoo tasks:
- AI reads the PRD from a file or pasted text
- Calls
odoo_list_projectsto find the target project - Calls
odoo_get_project_stagesto understand the workflow - Decomposes the PRD into a task hierarchy:
- Top-level tasks = epics / feature areas
- Subtasks = individual work items
- Creates epics via
odoo_create_task, getting back IDs - Creates subtasks via
odoo_create_tasks_batchwithparent_idlinking - Tags are auto-created if they don't exist
Supported Odoo Versions
Works with any Odoo version that supports XML-RPC (12.0+):
- Odoo Community
- Odoo Enterprise
- Odoo.sh
- Odoo Online (SaaS)
Troubleshooting
Connection fails
- Verify your credentials by logging into Odoo manually
- Check that your API key is valid (regenerate if needed)
- Ensure
ODOO_URLdoesn't have a trailing slash or/odoopath - Verify the database name matches exactly
SSL errors
SSL certificates are handled automatically via the bundled certifi package. If you're behind a corporate proxy or firewall that intercepts SSL, add these environment variables to your config:
"env": {
"ODOO_URL": "...",
"ODOO_DB": "...",
"ODOO_USERNAME": "...",
"ODOO_API_KEY": "...",
"SSL_CERT_FILE": "/path/to/your/corporate-ca-bundle.pem",
"REQUESTS_CA_BUNDLE": "/path/to/your/corporate-ca-bundle.pem"
}
To find your system's certificate bundle:
python3 -c "import certifi; print(certifi.where())"
uvx errors
Verify uvx is installed and get the correct path:
which uvx
Use that exact path in your config.
Environment Variables
| Variable | Required | Description |
|---|---|---|
ODOO_URL |
Yes | Your Odoo instance URL (base URL only) |
ODOO_DB |
Yes | Database name |
ODOO_USERNAME |
Yes | Login email |
ODOO_API_KEY |
Yes | API key (not password) |
SSL_CERT_FILE |
No | Path to CA certificate bundle (only for corporate proxies) |
REQUESTS_CA_BUNDLE |
No | Path to CA certificate bundle (only for corporate proxies) |
LOG_LEVEL |
No | DEBUG, INFO (default), WARNING, ERROR |
License
MIT
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。