Feature Request: Add Kilo Gateway (kilo.ai) as a native OpenAI-compatible provider
Provider Overview
Kilo Gateway (kilo.ai/gateway) is a universal AI inference API that exposes 500+ models (Anthropic, OpenAI, Google, xAI, Mistral, MiniMax, Z.ai, etc.) through a single OpenAI-compatible endpoint at https://api.kilo.ai/api/gateway. It uses standard Bearer authentication and supports the provider/model naming convention (similar to OpenRouter).
- Docs: https://kilo.ai/docs/gateway
- API:
https://api.kilo.ai/api/gateway
- Auth: Bearer token (
KILO_API_KEY)
- Model routing:
anthropic/claude-sonnet-4.5, openai/gpt-5.5, google/gemini-3.1-pro, etc.
Motivation
Kilo Gateway has become a solid inference option for accessing many models through one key, especially for users who want to route across multiple underlying providers without managing individual API keys. Adding native support would make GoModel work out-of-the-box with Kilo for chat completions, streaming, and tool calling — all of which are already OpenAI-compatible on their end.
Implementation Notes
- Kilo is fully OpenAI-compatible (
/chat/completions, SSE streaming, tool calling), so this should be implementable as a thin provider wrapping the shared OpenAI transport — similar to how OpenRouter is handled.
- Model IDs use the
provider/model format and should be passed upstream unchanged.
My Contribution Offer & Constraint
I'm happy to implement this myself and open a PR (registration, factory, tests, example config, docs). However, I need to flag one limitation:
I currently only use Kilo's free inference tier for personal usage. I do not have a paid subscription or credits on the platform. Therefore golden JSON can only be provided to some extent, if this is even necessary!
Let me know how you'd like to proceed!
Feature Request: Add Kilo Gateway (kilo.ai) as a native OpenAI-compatible provider
Provider Overview
Kilo Gateway (kilo.ai/gateway) is a universal AI inference API that exposes 500+ models (Anthropic, OpenAI, Google, xAI, Mistral, MiniMax, Z.ai, etc.) through a single OpenAI-compatible endpoint at
https://api.kilo.ai/api/gateway. It uses standard Bearer authentication and supports theprovider/modelnaming convention (similar to OpenRouter).https://api.kilo.ai/api/gatewayKILO_API_KEY)anthropic/claude-sonnet-4.5,openai/gpt-5.5,google/gemini-3.1-pro, etc.Motivation
Kilo Gateway has become a solid inference option for accessing many models through one key, especially for users who want to route across multiple underlying providers without managing individual API keys. Adding native support would make GoModel work out-of-the-box with Kilo for chat completions, streaming, and tool calling — all of which are already OpenAI-compatible on their end.
Implementation Notes
/chat/completions, SSE streaming, tool calling), so this should be implementable as a thin provider wrapping the shared OpenAI transport — similar to how OpenRouter is handled.provider/modelformat and should be passed upstream unchanged.My Contribution Offer & Constraint
I'm happy to implement this myself and open a PR (registration, factory, tests, example config, docs). However, I need to flag one limitation:
I currently only use Kilo's free inference tier for personal usage. I do not have a paid subscription or credits on the platform. Therefore golden JSON can only be provided to some extent, if this is even necessary!
Let me know how you'd like to proceed!