Jev Model - Typed AI decisions via API, playground, and remote MCP
by•
Independent playground, hosted HTTP API, and remote MCP for TypeSafe AI’s Jev decision model — not affiliated with or endorsed by TypeSafe AI. Send application state + typed questions; get choice, score, or calibrated yes/no probabilities your code can branch on — not chat. MCP tools: jev_decide, jev_noul, jev_choice, jev_score.
Replies
Hi Product Hunt! I built jevmodel.org as an independent playground, hosted API, and remote MCP for TypeSafe AI's Jev decision model — not affiliated with or endorsed by TypeSafe. Instead of parsing LLM text for routing and guardrails, you send state + typed questions and get labels, scores, or yes/no probabilities your software can act on. Would love feedback on the MCP tools and docs. Disclaimer: at high volume, buying direct from TypeSafe or a major gateway is usually cheaper.