---
title: "Self-host the AI assistant with your own model"
url: "https://www.openstatus.dev/changelog/bring-your-own-ai-provider"
date: "2026-07-03"
author: "Moulik Aggarwal"
description: "Self-hosted openstatus can now run the AI assistant on any OpenAI-compatible model — NVIDIA NIM, vLLM, Ollama, and more."
category: "integrations"
image: "https://www.openstatus.dev/assets/changelog/bring-your-own-ai-provider.png"
---

Running openstatus on your own infrastructure? The in-dashboard AI assistant no longer depends on a hosted gateway — you can now power it with **your own model provider**.

Point it at any **OpenAI-compatible** endpoint ([NVIDIA NIM](https://build.nvidia.com/), vLLM, Ollama, OpenRouter, or a private gateway) by setting three environment variables in your `.env.docker`:

```env
AI_BASE_URL=https://integrate.api.nvidia.com/v1
AI_API_KEY=nvapi-xxxxx
AI_MODEL=meta/llama-3.1-70b-instruct
```

The model and API key stay entirely on your own infrastructure, and `AI_API_KEY` is optional for keyless local gateways like Ollama. The only requirement is that the model supports tool calling, which the assistant uses to manage your monitors and status pages.

See the [self-hosting guide](/docs/guides/self-hosting-openstatus) for the full configuration.