by bnsd55 · Agent Tool · ★ 70
jevmlx Typed decisions from a local model on Apple Silicon. One batched forward pass, every field at once, with a probability per field. Asking an LLM for JSON means parsing what it wrote, fixing what drifted, and retrying until it parses. jevmlx takes a schema of booleans, enums, and multi-selects, scores every allowed answer for every field in one forward pass, and assembles the JSON itself. Nothing is generated token by token: the output is valid by construction and every field carries a probability.
| Stars | 70 |
| Forks | 8 |
| Language | Python |
| Category | Agent Tool |
| License | MIT |
| Quality Score | 67.8233791266676/100 |
| Open Issues | 7 |
| Last Updated | 2026-10-14 |
| Created | 2026-09-17 |
| Platforms | python |
| Est. Tokens | ~15k |
These tools work well together with jevmlx for enhanced workflows:
Explore other popular agent tool tools:
jevmlx is Jev-style parallel constrained decisions for any MLX model on Apple Silicon. Typed, schema-valid JSON in one forward pass.. It is categorized as a Agent Tool with 70 GitHub stars.
jevmlx is primarily written in Python. It covers topics such as apple-silicon, jev, local-llm.
You can find installation instructions and usage details in the jevmlx GitHub repository at github.com/bnsd55/jevmlx. The project has 70 stars and 8 forks, indicating an active community.
jevmlx is released under the MIT license, making it free to use and modify according to the license terms.