Show HN: 3-step proxy to use agent mode with third party models in Copilot

3 months ago 2

OpenRouter provides access to various AI models from OpenAI, Anthropic, and others. However, GitHub Copilot's Agent Mode requires models to support function calling/tools, but OpenRouter's API doesn't announce tool support for its models.

This prevents using powerful models like Claude, GPT-4, and others through OpenRouter with Copilot's advanced Agent Mode features.

copilot-ollama creates a local proxy chain that:

🔄 Forwards requests to OpenRouter while preserving tool support
🛠️ Makes OpenRouter models compatible with Copilot's Ollama integration
🚀 Enables Agent Mode with any OpenRouter model
🔧 Uses LiteLLM for OpenAI-compatible proxying
🔗 Uses oai2ollama for Ollama compatibility

uv package manager
OpenRouter API key (get one here)
VSCode with GitHub Copilot extension

Clone and navigate to the project

git clone https://github.com/bascodes/copilot-ollama.git cd copilot-ollama
Set your OpenRouter API key

export OPENROUTER_API_KEY=your_openrouter_api_key_here
Start the proxy servers
Configure VSCode
- Open VSCode settings
- Set github.copilot.chat.byok.ollamaEndpoint to http://localhost:11434
- Click "Manage Models" → Select "Ollama"
Start coding! 🎉 Your OpenRouter models are now available in Copilot Agent Mode.

Edit config.yaml to add or modify available models:

# This section defines the models that your local proxy will advertise model_list: - model_name: kimi-k2 # Name that appears in VSCode litellm_params: model: openrouter/moonshotai/kimi-k2 # Actual OpenRouter model - model_name: claude-3-sonnet litellm_params: model: openrouter/anthropic/claude-3-sonnet - model_name: gpt-4-turbo litellm_params: model: openrouter/openai/gpt-4-turbo

Popular OpenRouter Models

Here are some recommended models to add:

Model Name OpenRouter Path Description

claude-3-sonnet	openrouter/anthropic/claude-3-sonnet	Excellent for code generation
gpt-4-turbo	openrouter/openai/gpt-4-turbo	Latest GPT-4 with improved performance
mixtral-8x7b	openrouter/mistralai/mixtral-8x7b-instruct	Fast and capable open-source model
llama-3-70b	openrouter/meta-llama/llama-3-70b-instruct	Meta's powerful open model

graph LR A[VSCode Copilot] --> B[oai2ollama<br/>:11434] B --> C[LiteLLM Proxy<br/>:4000] C --> D[OpenRouter API] D --> E[AI Models<br/>Claude, GPT-4, etc.]

VSCode Copilot sends requests to what it thinks is an Ollama server
oai2ollama translates Ollama API calls to OpenAI format
LiteLLM proxies OpenAI-compatible requests to OpenRouter
OpenRouter routes to the actual AI model providers
Tool/function calling capabilities are preserved throughout the chain

We welcome contributions! Here's how you can help:

🐛 Report bugs by opening an issue
💡 Suggest features or improvements
📖 Improve documentation
🔧 Submit pull requests

# Clone the repo git clone https://github.com/bascodes/copilot-ollama.git cd copilot-ollama # Install dependencies uv sync # Make your changes and test export OPENROUTER_API_KEY=your_key ./run.sh

This project is licensed under the MIT License - see the LICENSE file for details.