The tool is a free, MIT-licensed CLI now in beta. It entered beta on October 5, 2026 and works with Claude Code, Claude Desktop, Codex, ChatGPT Desktop, OpenCode (version 2+), and Pi (version 0.80.8+). Available models include Kimi K3, GLM 5.3, MiniMax M3, and DeepSeek V4.1 Flash. The router decides once per session, and routing once per session is intended to preserve prompt caching. If a user supplies an Anthropic key, it routes between Opus 5.5 and GLM 5.3. Without one, it routes between GLM 5.3 and GLM 5.3 Flash. Together pitches the tool at a specific cost problem: it says engineering orgs spend tens of thousands to millions of dollars monthly on closed models. The company already serves the largest share of OpenRouter tokens for DeepSeek V4.1 Flash (40.8%), GLM 5.3 Flash (28.2%), and Kimi K3 (23.1%). Outside reviewers note that actual savings depend on routing, token usage, caching, and model choice. For founders, the tool separates the choice of coding agent from the choice of model vendor. That gives teams a cheap way to test open models on real coding work. It also puts pressure on closed labs, whose own agents are now carrying traffic to competing models.