jev-visual
hr98w
Local Jev-like visual inference experiment on Apple Silicon Mac. Scores and classifies single images across multiple questions with 3 playable game demos.
- License
- MIT
- GitHub Stars
- 298
- Source reviewed
- —
Where Jev makes a decision
Reuses multimodal visual context to score candidate answers directly from logits without autoregressive generation.
What this project offers
Brings Jev-style single-pass multi-question scoring to local edge vision on Apple Silicon.
Review scope
Open-source experiment running local multimodal Jev-style scoring on Apple Silicon (MLX), not official TypeSafe cloud service.
Sources and implementation
Related projects
jev
dannote · SDK & Decision Frameworks
Integrates Jev as an asynchronous Elixir/OTP process whose replies are handled with GenServer pattern matching.
zod-jev
jomatsu · SDK & Decision Frameworks
Adds semantic rules to Zod validation, such as checking whether text matches a description or contains personal information.
daf-jev
docxology · SDK & Decision Frameworks
A Python toolkit for Jev requests, batch evaluation, calibration and MCP access.
jev-starter
hamakyo · SDK & Decision Frameworks
TypeScript patterns for thresholds, fallbacks, human review and evaluation on top of the TypeSafe SDK.