Martian Documentation
As we perform research, we often build new solutions and tools to meet our workflows. Martian makes some of these solutions available externally.
Additional tooling is available for our partners. Please log in to access additional documentation.
Gateway
The Martian Gateway provides unified access to 200+ AI models through a single API, so you can pick the most effective model for your use case while maintaining compatibility with popular frameworks.
- Access to 200+ models from leading providers
- Compatibility with OpenAI and Anthropic APIs
For example, simply populate the model field when making requests:
curl https://api.withmartian.com/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $MARTIAN_API_KEY" \
-d '{
"model": "openai/gpt-4.1-nano",
"messages": [
{
"role": "user",
"content": "What is Olympus Mons?"
}
]
}'
Track your usage in the Martian Dashboard: real-time usage, model performance, and request history.
ARES
ARES (Agentic Research and Evaluation Suite) is an RL-first framework for training and evaluating LLM agents, especially coding agents. It treats LLM requests as observations and LLM responses as actions, so you can train the LLM within the agent—not the whole agent as a black box.
- RL-first design with async, scalable environments
- Full documentation on Read the Docs (Core Concepts, How It Works)
- Open source: github.com/withmartian/ares
Getting Help
- Support: support@withmartian.com