Connect AI securely to Odoo, your ERP, databases and internal business applications — without giving up control of your data. Deploy entirely inside your own infrastructure: on-premise, private cloud, or dedicated AI hardware.
A Model Context Protocol (MCP) server is a gateway that lets AI assistants talk to your business systems in a controlled and auditable way. Instead of giving AI unrestricted access to your ERP, databases or APIs, the MCP server exposes only the tools and information you approve — governed by your organisation's own security policy.
The AI never talks directly to your business systems — every request passes through your MCP server, where it is checked, scoped and recorded.
Most AI integrations require sending business information out to an external platform. OdooBridge Enterprise is built the other way round: it brings AI to your data, so your organisation decides exactly where AI runs.
A server in your own building, on your own network.
Inside existing infrastructure, alongside the systems it serves.
Your own AWS, Azure or Google Cloud tenancy — your account, your controls.
Pre-configured hardware, ready to connect to Odoo.
Air-gapped environments are supported too — see deployment options. Just evaluating for now? Start with the hosted trial against your own Odoo — then bring the same tools inside your infrastructure.
Enterprise deployments layer several controls, designed for organisations handling sensitive business information.
Runs entirely inside your own infrastructure, with no mandatory dependency on our cloud.
Enterprise deployments ship as protected application packages — designed to safeguard intellectual property while giving you complete operational control.
Encrypted channels between the AI, the MCP server and your business systems.
Different departments get different AI permissions. Finance AI can't reach HR; sales AI can't reach payroll. Every tool is individually controlled.
Every action is recorded: who did it, which records were touched, and each changed field's old and new value. Review it in the built-in web viewer, filter by user or record, export CSV — and restore any previous value in one approved step. The log lives on your infrastructure and never leaves it. Odoo itself keeps no comparable trail.
Deployments can be licensed to dedicated infrastructure, helping ensure software integrity and authorised usage.
Unlike many AI services, your organisation decides how the whole system behaves — not the vendor.
This is where OdooBridge stands alone. Neither Odoo Community nor Odoo Enterprise keeps a real who-changed-what history — governance teams reach for a third-party module and still can't see it per AI action. Because every request flows through one governed gateway, OdooBridge records the lot — and on your deployment, the trail never leaves your infrastructure.
Every tool call is logged with the signed-in Odoo user, the tool, the model and record, and the outcome — and each changed field is stored as an exact old → new value, not merely a field name.
A /audit timeline with old→new diffs and a /reports usage dashboard — activity by user, tool, model, outcome and day, with error rates and latency. Filter, drill in, export CSV. Nothing installed in Odoo.
Who can review the trail is decided by each person's role in your Odoo — administrators and managers see the whole organisation; everyone else, only their own actions. Enforced server-side, never spoofable from a client.
Beyond the queryable store, every event streams to a JSONL file and an optional webhook — forward it to Splunk, OpenSearch, Loki or any SIEM you already run.
On a private deployment the audit database, the dashboards and the full before/after values all live on your own machines. It is your compliance record, under your control — it never touches our servers.
Because the prior value is recorded, restoring it is a single previewed, approved change — deliberately safer than a blind “undo”, and itself written to the trail.
Choose today's best AI — or tomorrow's. Your MCP server stays the same; only the model changes.
Because the governed tool layer lives in your MCP server rather than in the model, switching providers doesn't mean rebuilding your integration, re-approving your tools, or renegotiating where your data goes.
Deploy inside your own AWS, Azure or Google Cloud environment. Suits enterprises that need scale while keeping control of the environment.
Deploy entirely inside your network — the usual choice for regulated industries and strict data-residency requirements.
A pre-configured enterprise AI server with the Odoo AI bridge ready to use. Plug in, connect Odoo, start working.
For environments where connectivity is restricted. The AI, MCP server and business systems all stay inside your secure environment.
Instead of copying information between systems, your AI interacts with Odoo through your MCP server — within the permissions you granted it.
Not every team starts here. The same tools and safety rails come in four forms — you can evaluate on the lightest and grow into a governed deployment when you're ready.
Zero install, works with ChatGPT/Claude/Cursor. Point our hosted server at your Odoo to evaluate.
An app that runs the bridge on your machine; your Odoo credentials live in your OS keychain.
Run it yourself with Claude, Cursor or Gemini — credentials never leave your network.
A private MCP server inside your infrastructure with role-based access, audit and policy.
See all four options compared side by side, or start with the hosted trial against your own Odoo.
| Typical AI integration | OdooBridge Enterprise |
|---|---|
| Public-cloud dependent | Runs inside your infrastructure |
| Limited control | Full enterprise control |
| Governance is difficult | Complete audit trail |
| Odoo keeps no change history | Every action logged, with before & after |
| No visibility into AI actions | Built-in /audit & /reports dashboards |
| Vendor lock-in | AI provider independent |
| Generic, broad AI access | Controlled, approved MCP tools |
| Shared infrastructure | Dedicated deployment |
How we talk about security. We describe what the architecture is designed to do, not absolutes. OdooBridge Enterprise supports fully local and air-gapped deployments, encrypted communication, controlled access and protected application packaging — and your organisation controls where data is processed. What any specific deployment guarantees depends on how it is configured and operated, which is exactly what we scope with you before installation.
Private infrastructure. Private AI. Private MCP server. Enterprise security architecture. Built for organisations that want the power of AI while keeping ownership of their business data.