GitHub · by ikermoel
Typed, calibrated decisions from any open-weights model, on your own GPU.
A Python library that answers typed questions from the next-token distribution of an open-weights model, restricted to your options, in one forward pass. It runs on Hugging Face Transformers and vLLM, with a live demo on Hugging Face Spaces and benchmarks in the repository.
Open the source
Tony Dinh
@tdinh_me
Just trying out Jev, I made a Chrome extension that: - Listens to your YouTube audio (optional) - Detects if it gets to a sponsor segment - Skips it ➡️➡️➡️ - All in real-time while costing ~$0.005 per video Prototype project, BYOK, open-source: github.com/trungdq88/yout…
GitHubTools and apps
An MCP server with claim checks, screening and ranking built on Jev.
Joey Kudish
70
GitHubTools and apps
An MCP connector that gives your agent direct access to Jev.
Mostafa
62