Technical

How to Use a Local LLM (Ollama) with Odoo

Keep your data in-house by running AI on your own server. Here is how to connect Ollama, vLLM or another OpenAI-compatible local model to Odoo with AI Base, with the same caps and usage log.

KalvyAI Engineering · · 2 min read

AI provider form in Odoo with provider, model, Fetch Models, API key, base URL, temperature, max output tokens, timeout and monthly token cap

Some companies cannot send invoices, contracts or customer data to a cloud AI provider. A local LLM solves that: the model runs on your own server and your data never leaves your network. With the free AI Base module, Odoo can use a local model just like OpenAI or Claude, with the same monthly caps and usage log.

What you need

  • A model server with an OpenAI-compatible chat-completions endpoint, such as Ollama, vLLM or a self-hosted NVIDIA NIM. Ollama serves one at http://<your-server>:11434/v1.
  • A network route from your Odoo server to that model server.
  • AI Base installed in Odoo 19 or Odoo 20.

Connect a local LLM to Odoo

  1. Go to AI Providers > Providers > New.
  2. Choose a provider type that uses Chat Completions, such as Mistral AI or NVIDIA NIM.
  3. Set the Base URL to your server’s endpoint, for example http://10.0.0.5:11434/v1.
  4. Type the model name your server exposes, exactly as it is installed there.
  5. If the form asks for an API key and your server does not use one, enter any placeholder text.
  6. Click Test Connection to confirm the model answers.
Watch out: "localhost" means the Odoo server itself. If Odoo runs on Odoo.sh or another host, it cannot reach a model on your laptop; use an address the Odoo server can actually reach.

What works the same as cloud AI

  • Every module built on AI Base can use the local model with no code changes.
  • Monthly token caps and the full usage log still apply, which helps you plan hardware capacity.
  • You can keep a cloud provider as a backup for when the local server is busy.

Native PDF reading is available only on OpenAI, Claude and Gemini. With a local model, image support depends on the model you run.

FAQ

Does my data leave my network with a local LLM?

No. AI Base sends each call straight from your Odoo server to the Base URL you set, and stores nothing outside your Odoo.

Can I mix local and cloud models?

Yes. Add as many providers as you like, local and cloud, each with its own cap. See the full AI Base guide for multi-provider setup.

Get started

Get AI Base free from the kalvyai Store, then follow the complete AI Base user guide for every setting, user roles and usage reports.

  • Odoo
  • Ollama
  • Local LLM
  • Self-hosted AI
  • AI Base

KalvyAI Engineering · Engineering team

The engineers who build, connect and fix business systems at KalvyAI: Odoo, ERPNext, Shopify and AI agents that work inside them with approvals, budgets and an audit trail.

Comments

No comments yet. Questions and corrections are welcome.

Rate this post or leave a comment

Your rating (optional)
Not rated

Comments appear after an editor approves them. See our privacy notice.