A Model Context Protocol (MCP) server for Odoo. Give Claude, ChatGPT, and other AI agents controlled access to your ERP — with permissions, audit logs, and no risky direct database access. Pre-launch: join the waitlist for early access.
AI agents are becoming useful at real work — answering questions over your data, drafting records, and triggering actions. But most of them can’t safely touch your ERP. Odoo MCP closes that gap. It is a Model Context Protocol (MCP) server purpose-built for Odoo, so AI agents and large language models can read and act on your business data through a controlled, permissioned interface instead of raw database access or brittle scripts.
We are in pre-launch. The fastest way to get Odoo MCP is to join the waitlist — early-access members get priority onboarding, setup support, and direct input on the roadmap. Join the waitlist.
What is Odoo MCP?
Odoo MCP is a connector that speaks the Model Context Protocol — the open standard for letting AI assistants discover and use external tools and data. On one side it talks to your Odoo instance through its standard API. On the other side it exposes a set of safe, well-defined tools and resources that any MCP-compatible AI client can call. The result: an AI agent can answer “which invoices are overdue this week?” or “create a draft quotation for this customer” without anyone hand-coding a one-off integration.
Why not just give the AI my Odoo login or database?
Because that is how data leaks and accidents happen. Direct database access bypasses Odoo’s own security model, and sharing a full user login gives an agent far more power than any single task needs. Odoo MCP sits in between: it scopes what the agent can see and do, enforces permissions, and logs every call. You get the productivity without handing over the keys to everything.
How Odoo MCP works
The flow is intentionally simple:
- AI client — Claude Desktop, an IDE assistant, a custom agent, or any MCP-compatible app — sends a request in plain language.
- Odoo MCP server translates that into safe, validated operations: reading records, searching, summarising, or (where you allow it) creating and updating records.
- Permission layer checks what this agent and this user are actually allowed to touch, respecting Odoo access rights and record rules.
- Audit log records what was requested and what was done, so nothing happens in the dark.
- Odoo returns the result, and the AI presents it back to the user in context.
Why this matters now
Every team running Odoo already has the data that makes AI useful — customers, orders, inventory, invoices, projects. What’s missing is a safe bridge. Custom API scripts age badly, break on upgrades, and each one is a new thing to maintain. MCP standardises the bridge, so the same Odoo connection works across many AI tools and many use cases. Build (or buy) it once; reuse it everywhere.
What you can do with Odoo MCP
- Conversational ERP — ask questions about sales, stock, or finance in natural language and get answers grounded in live Odoo data.
- Drafting and data entry — let an agent prepare quotations, leads, tasks, or product records for human review.
- Reporting on demand — pull a quick summary without building a dashboard for every question.
- Operational copilots — give support, sales, or warehouse teams an assistant that actually knows the company’s data.
See more in Odoo MCP use cases, learn the fundamentals in What is Odoo MCP, and read the case for adopting it in Why Odoo MCP.
Built with security first
Odoo MCP is designed around least privilege. You decide which models and fields are reachable, whether an agent can only read or also write, and which users an agent acts on behalf of. Every request is auditable. This is the difference between an experiment and something you can put in front of a real team.
Frequently asked questions
Is Odoo MCP available today?
Not yet — we are in pre-launch. Join the waitlist to get early access, priority onboarding, and updates as we ship.
Which AI tools will it work with?
Any client that supports the Model Context Protocol — including Claude Desktop and a growing list of MCP-compatible agents, IDEs, and custom applications. Because MCP is an open standard, you are not locked into one vendor.
Will it work with my Odoo version?
We are targeting current and recent Odoo releases and both self-hosted and cloud setups. Tell us your version when you join the waitlist so we can prioritise compatibility.
Does it replace Odoo’s own AI features?
No. Odoo MCP is about giving external AI agents safe access to your Odoo data and actions, complementing whatever you run inside Odoo. It is the secure bridge, not a chatbot.
How is this different from writing my own API integration?
A custom integration is one-off code you build, secure, and maintain for every tool and every task. Odoo MCP gives you one standardised, permissioned, audited interface that works across many AI clients and many use cases — far less to build and far less to break on upgrade.
What does early access include?
Priority onboarding, hands-on setup help, influence over the roadmap, and first access before public launch. Join the waitlist.
Be first to connect AI to your Odoo — safely. Join the Odoo MCP waitlist and we’ll reach out before launch.
Get in touch
Want early access to Odoo MCP? Join the waitlist below. Tell us your Odoo version and what you’d connect an AI agent to first — we’ll reach out before public launch with setup help and priority onboarding.
Be first to connect AI agents to your Odoo, securely. Odoo MCP is in pre-launch, and waitlist members get early access before public release. Reserve your spot, tell us your Odoo version and first use case, and get priority onboarding plus a direct line to the roadmap. No spam — just launch updates and your invite when it’s ready.
📄 Odoo MCP technical overview (for devel…