[FEATURE]: OpenAI Compatible Providers

[smaEti]

Feature hasn't been suggested before.

• I have verified this feature I'm about to request hasn't been suggested before.

Describe the enhancement you want to request

Currently, Kilo Code CLI seems to be limited to specific hardcoded providers. This restricts users from leveraging:

1. Local Models: Running local LLMs via Ollama, LocalAI, or LM Studio for privacy and cost reasons.
2. Alternative Providers: Using high-speed or specialized providers that use the OpenAI-standard API (like Groq, Together AI, DeepSeek, or OpenRouter).
3. Enterprise Gateways: Corporate proxy endpoints that mimic the OpenAI schema.

I would like to see a new provider type in the configuration (e.g., openai-compatible or custom). This provider should allow users to specify:

• base_url: The custom endpoint (e.g., http://localhost:11434/v1)
• api_key: A field for the key (or a dummy string for local providers).
• model: The specific model string to be passed to the API.

Currently, users are forced to use the official OpenAI/Anthropic keys, which might not be viable for developers working on sensitive codebases who prefer local inference.

If I have missed an existing setting for this, please let me know!

Thank you for your help!

[github-actions]

This issue might be a duplicate of an existing feature request. Please check:

• [FEATURE]: Add Ollama/Local Model Provider Support #154: [FEATURE]: Add Ollama/Local Model Provider Support - This feature has already been implemented with support for Ollama and OpenAI-compatible endpoints, allowing users to configure custom base URLs and API keys for providers like Groq, Together AI, DeepSeek, and others using the OpenAI-standard API.

[Nomadcxx]

The fix here is having kilo as a provider in opencode: anomalyco/opencode#13765

[smaEti]

Thanks for the follow up! To clarify my request: I’m not looking for a specific new provider to be integrated. Instead, I’d like the ability to use OpenAI-compatible custom endpoints, similar to how the VS Code extension currently functions.

Many users (myself included) rely on local models via Ollama, LM Studio, or vLLM, as well as third-party providers. These all use the OpenAI API specification.

To support this, we should have an option to manually configure:

Base URL: (e.g., http://localhost:11434/v1)
API Key: (Custom or placeholder key)
Model ID: (A text field to manually type the model name)
Bringing this feature over from the VS Code extension would provide a lot more flexibility for those of us running local LLMs or using non-standard providers.