awesome-jev
Launch what you built on Jev, TypeSafe AI's System One decision model: a weekly board where makers launch and people vote, over an index of public Jev examples
What is awesome-jev?
Launch what you built on Jev, TypeSafe AI's System One decision model: a weekly board where makers launch and people vote, over an index of public Jev examples by decision pattern, with source citations, dated link checks and call-site text checks; runtime and performance are not independently tested here. EN/中文, JSON schema, platform matrix.
Overview of awesome-jev
What it is
Jev is a decision model from TypeSafe AI. It does not write text — it accepts state plus typed questions and returns typed answers with calibrated confidence. This repository indexes 1207 public examples of using Jev, organized by the decision being made. It is not the product, not an SDK, not affiliated with TypeSafe AI, and not a recommendation. Inclusion is a source record, not a runtime or performance endorsement.
Key features
- Three primitives: Choice, Score, and Noul. Every decision pattern is built from them.
- Input is text only — string, JSON object, or array of text.
- Context is 64k tokens per request, 32k for the state plus the longest question. Output tokens are free.
- No published weights, so it cannot be run locally.
- 1207 public resources indexed by decision pattern, with source citations, dated link checks, and scheduled call-site text checks.
- Runtime and performance are not independently tested here.
Use cases
- Tool selection (231 entries)
- Context compaction (35 entries)
- Safety gating (140 entries)
- Output validation (135 entries)
- Retry control (7 entries)
- Human escalation (69 entries)
- Model routing (45 entries)
- Speculative fan-out (33 entries)
- Search and ranking (64 entries)
- Structured extraction (17 entries)
- Classification (120 entries)
- ML feature extraction (8 entries)
- Document triage (20 entries)
- Support triage (9 entries)
- Content scoring (166 entries)
- Recommendation (1 entry)
- Overview (453 entries)
- Also includes negative results from independent measurements, such as Hermes Agent compaction evaluation and worldmonitor news threat classification.
Getting started
See the repository README for quickstart commands (Python, JS, cURL examples are referenced but not reproduced here).
License and pricing
License is not stated (NOASSERTION). The project is open source.
Based on the project's public README and repository metadata.