mcp-server-odoo

What it is and what it does

mcp-server-odoo is a Model Context Protocol server that runs locally on your machine and acts as a bridge between AI assistants (Claude, GitHub Copilot, etc.) and an Odoo ERP instance. It translates natural-language requests from the AI into Odoo API calls, allowing you to search records, create entries, update data, delete records, count matches, inspect model schemas, and perform server-side aggregations—all through conversational prompts rather than manual API work.

The package supports two modes: standard mode (production-ready, requires the Odoo MCP module installed on your Odoo instance for security and access control) and YOLO mode (quick testing and demos, works with any Odoo version with XML-RPC enabled). It handles authentication via API key or username/password, supports multi-language responses, and includes smart pagination and field selection. Configuration is environment-variable-driven, making it straightforward to integrate into Claude Desktop, Cursor, VS Code with Copilot, or other MCP-compatible clients.

Use it for:

• Query customer or product data from Odoo using natural language in Claude, without writing SQL or API calls.
• Create purchase orders, invoices, or other records through conversational prompts, with field validation handled server-side.
• Aggregate sales data (sum, count, group by) for reporting or analysis without downloading raw rows.
• Inspect Odoo model schemas to understand available fields and relationships before querying or updating.
• Automate workflow actions (post invoice, confirm sales order) via Claude by invoking public Odoo methods.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

An MCP server that bridges AI assistants like Claude to Odoo ERP systems, enabling natural-language queries and operations on business data through search, create, update, delete, and aggregation tools.

Yes, if you use Claude or another MCP-compatible AI assistant and need programmatic access to an Odoo instance. Install friction is low, maintenance is active, and there are no known vulnerabilities. The copyleft license (MPL-2.0) is not a blocker for internal use. Start with YOLO mode for testing; move to standard mode with the Odoo MCP module for production security.

Install

mcp-server-odoo on PyPI

pip install mcp-server-odoo

uv add mcp-server-odoo

poetry add mcp-server-odoo

Installing mcp-server-odoo

Before you install

Low friction: pure Python wheel with four lightweight runtime dependencies (mcp, pydantic, pydantic-settings, python-dotenv). Active maintenance with recent commits and steady downloads. Requires Python 3.10+.

License in practice

MPL-2.0 copyleft license: derivative works must be licensed under MPL-2.0, and source code changes must be made available. Acceptable for internal tools and integrations; review required if redistributing modified versions.

Quickstart

# Install via uvx (recommended for MCP)
uvx mcp-server-odoo

# Or via pip
pip install mcp-server-odoo

# Configure in Claude Desktop or other MCP client:
# Set ODOO_URL and ODOO_API_KEY environment variables,
# then the server exposes search, create, update, delete tools.

Requires Python 3.10 or higher. Requires an Odoo instance (version 16.0+ for standard mode with MCP module, or any version with XML-RPC enabled for YOLO mode). MCP client (Claude Desktop, Claude Code, Cursor, etc.) must be configured to load the server.

Verify before relying

• Whether the package's smart field selection and LLM-optimized output formatting actually reduce token usage or improve response quality in practice.
• Performance characteristics when querying large Odoo datasets or handling concurrent requests from multiple AI assistants.
• Specific Odoo versions and configurations known to work reliably in production (beyond the stated 16.0+ requirement).

Package facts

Evidence: mcp_server_odoo-0.7.1-py3-none-any.whl

Keywords: ai, anthropic, claude, erp, llm, mcp, model-context-protocol, odoo

Tags

More Python Modules packages

Further reading