Use supported local model providers

Version 3.0 supports local and self-hosted models through provider APIs rather than through the previous Python local server workflow.

Use one of the following options:

• Ollama, when your model is served by Ollama.

• OpenAI-compatible API, when your local or self-hosted runtime exposes an OpenAI-compatible /v1 API.

Local and self-hosted providers keep HTTP traffic on infrastructure you control. The extension still sends the selected request, response, prompt, and chat messages to the endpoint you configure.

After configuring a local provider, run Test request from Provider settings before scanning traffic or using AI Chat.