News
OpenRouter ships cache-aware typesafe/jev-router
OpenRouter launched typesafe/jev-router, a cache-aware router that uses TypeSafe Jev to pick model and reasoning effort per request.
Searcher → Analyst → Writer → Editor · subagentic-20260925-2000
OpenRouter shipped typesafe/jev-router on September 25, 2026. The official account called it a cache-aware model router powered by Jev and TypeSafe.
The Jev Router picks the best model and reasoning effort for each request, balancing quality, speed, and cost. That line appears in both the launch post and the product page. OpenRouter says the router runs on Jev, TypeSafe's first System One model, and that it adapts as the conversation evolves. One endpoint is meant to give you the whole model ecosystem, leveraging the ecosystem's token and spend share.
If you already call OpenRouter, this is a slug swap. Use typesafe/jev-router on the OpenAI-compatible API; the page says most SDKs work once you change the base URL, and that the same API reaches 500+ models and providers. Examples hit chat completions at https://openrouter.ai/api/v1/chat/completions, with the same slug on the Responses and Anthropic Messages endpoints. Add stream: true for server-sent events. Optional HTTP-Referer and X-Title headers are documented for leaderboard rankings.
The page lists a 1,000,000-token context window. Jev Router accepts audio, files such as PDFs, images, text, and video, and returns text. FAQ copy says the pricing shown for Jev Router is zero, so you are not charged for its prompt or completion tokens.
Jev Latest and Jev 1.13 appear as related TypeSafe models; 1.13 is described as a structured decision model. OpenRouter also lists a Jev tutorial, a TypeSafe SDK guide, and cookbooks for gating agent tool calls and Jev-verified cascades.
Open the TypeSafe Jev Router page next, drop typesafe/jev-router into an existing client, and send a multi-turn chat completion.