Support for local models

[diro5t]

Describe the feature or problem you'd like to solve

Support of leveraging local models for cost, privacy and safety reasons is a must nowadays

Proposed solution

Please allow usage of Ollama, LM Studio or offer your own solution to accomlish this

Example prompts or workflows

No response

Additional context

No response

[examon]

It looks like this was addressed in v1.0.13. The CLI can use local model endpoints via the COPILOT_PROVIDER_BASE_URL environment variable -- Ollama, vLLM, and other OpenAI-compatible providers are supported, GitHub authentication is not required when using a custom provider, and fully air-gapped usage works with COPILOT_OFFLINE=true. If you believe this isn't fully resolved, feel free to reopen this issue.