
Free / Open-sourceBuild, simulate, evaluate and monitor self-improving AI agents

Selene 1 has officially shut down after its parent company Atla wound down operations and its domain began returning 404 errors. No active API or evaluation services remain accessible.
Direct outbound link disabled because atla-ai.com was unavailable or returning server errors during verification.
Verified active tools in AI API, ranked by current traffic and editorial score.

Free / Open-sourceBuild, simulate, evaluate and monitor self-improving AI agents


Google's rapid prototyping IDE and developer API workbench for building multimodal applications with Gemini models.


Free / Open-sourceAgent observability, evals and governance for AI platform teams

The data, evaluations and systems behind AI that has to be right.
One OpenAI-compatible gateway for adaptive LLM routing, guardrails, and spend control
AI gateway, observability and guardrails for production LLM apps
Selene 1 was developed by Atla as a specialized suite of AI evaluation models designed to benchmark and refine generative AI applications. It acted as an advanced LLM-as-a-Judge tool, enabling engineering teams to test prompt accuracy, secure LLM outputs, and run systematic quality checks before production deployment. Despite strong early interest in automated AI testing, the project was wound down following corporate restructuring and team acquisitions.
Atla's Selene 1 is permanently offline and its domain returns 404 errors. Check the top of this page for active AI evaluation and prompt testing platforms that can seamlessly replace it.
Free / Open-sourceBuild, simulate, evaluate and monitor self-improving AI agents
Google's rapid prototyping IDE and developer API workbench for building multimodal applications with Gemini models.
Free / Open-sourceAgent observability, evals and governance for AI platform teams
One OpenAI-compatible gateway for adaptive LLM routing, guardrails, and spend control