GitHub radar
SemIf: Semantic AI Decisions Without Text Generation
An independent open-source tool that reads typed option probabilities directly from a local model — no answer sentence, no JSON parsing loop, 5.21x faster than asking a chat model.
SemIf is an independent open-source tool that reproduces the interface pattern of TypeSafe's Jev: it reads typed option probabilities directly from a local model instead of generating an answer sentence. You supply a text state and a schema of questions — each question is a choice from a list or a yes/no. The model returns the probability of each option. No answer token is sampled, no JSON is parsed. On an RTX 3090, 21 binary decisions take 1.023 seconds, compared to 5.332 seconds for an autoregressive JSON array — 5.21x faster. Apple Silicon is supported via an MLX backend with serial prefix reuse and parallel shared-state decisions. A WebGPU browser demo runs without any installation. The project was formerly called OpenJev and is not affiliated with TypeSafe.
Why a vibe-coder should care
If your agent needs to classify incoming messages, route tickets, or check conditions, SemIf replaces a full chat model call with a direct probability read — faster and cheaper per decision. Runs locally, no cloud API required. For simple classification workloads it reaches close to Jev's accuracy on most tasks I tried.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Look at SemIf: https://github.com/TheoLeeCJ/SemIf — a tool for text classification without generation. Install it on my Mac using the MLX backend from the README, run the example from examples/decisions.jsonl and show me the results; ask me if you need a model path.
Runs on any Apple Silicon Mac via the MLX backend — no discrete GPU needed. On Windows or Linux, an NVIDIA GPU with 8+ GB VRAM is required for the 4B model.
Open on GitHub▌ More finds