Is there a way to use local AI from jetbrains plugin in Copilot CLI mode ?

[abenevaut]

🏷️ Discussion Type

Question

💬 Feature/Topic Area

JetBrains & Xcode

Body

I'm working with Copilot CLI in terminal with local Ollama model and that works perfectly, is there a way to do it from jetbrains IDE ? With the copilot plugin https://plugins.jetbrains.com/plugin/17718-github-copilot--your-ai-pair-programmer

To accomplish this on Copilot CLI, i use env. variables, maybe there is specific env. vars to do this inside the plugin ?
https://gist.github.com/abenevaut/2f0b39c6b0336e8382e86b6912334364

[shailesh-zx]

As of now, there is no official support for using a local AI model (like Ollama) directly inside GitHub Copilot CLI or the JetBrains Copilot plugin.

Copilot CLI is tightly integrated with GitHub’s hosted models, and it does not expose a configuration layer to swap the backend model via environment variables or plugin settings.

If you are already using Ollama locally, the closest alternatives would be:

• Using Ollama directly in the terminal alongside Copilot CLI (as separate tools)
• Using IDE plugins that support local models, such as Continue.dev or similar extensions
• Building a custom bridge using LSP or API proxy (advanced use case, not officially supported)

At the moment, Copilot itself does not provide a mechanism to route requests from JetBrains plugin or CLI to a local model backend.

[elzharaawy]

GitHub Copilot CLI uses GitHub's hosted AI models through your Copilot account and does not support directly using the local AI models provided by the JetBrains AI Assistant plugin.

Your options are:

Copilot CLI → Uses GitHub-hosted models only.
JetBrains AI Assistant → Can use supported local models within the IDE (depending on version and configuration).
For a local model in the terminal, use tools designed for local inference such as:
Ollama
LM Studio
Open WebUI

If your goal is to have a Copilot-like CLI experience with a local model, you can run a local LLM via Ollama and use CLI tools that connect to it, but Copilot CLI itself does not currently provide a setting to switch from GitHub-hosted models to JetBrains local models.

[abenevaut]

Copilot CLI → Uses GitHub-hosted models only.

No, let's refresh dataset.

https://docs.github.com/fr/copilot/concepts/agents/copilot-cli/about-copilot-cli#using-your-own-model-provider

[abenevaut]

Seems possible with jetbrains junie plugin and opencode via ACP

[AbrahamRdzSus]

Does anyone know how to contact GitHub support directly? The automation isn't working, and neither is the GitHub support email. I've even tried Copilot, a ticket, and Corrobo, and nothing has worked.