Automated README summary
Automatically summarized from the README on 31 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
A drop-in proxy for the Jev SDK that learns a small local model from Jev's own answers on your traffic, then answers most repeated classification requests itself while keeping agreement within a budget you set.
How it uses Jev
Jevstiller forwards requests to Jev until a task's student is trained and passes checks. It learns from Jev's full probability distributions, picks a routing threshold via a Clopper–Pearson bound on held-out calibration data, and answers locally only when confident; uncertain input and a random audit slice still go to Jev.
Primitives:choicenoul
Technique worth stealing
Choose the local-answer threshold with an exact finite-sample (Clopper–Pearson) bound on disagreement over all requests, then re-verify it forever on a random audit slice.
Try it
docker run -d -p 8080:8080 -v jevstiller-data:/data ghcr.io/tomerglick57/jevstiller, then set TYPESAFE_BASE_URL=http://localhost:8080.
Model judgments from README excerpts
Project signals use repository metadata, an existing summary and the first 700 README characters, not the full project. “Not found” means not detected in that material; it does not establish absence. These are the original judgments with recorded models and dates, not source audits or run results.
Models & evaluation dates
- Inclusion gate
jev-1.14.0- Project signals
jev-1.14.0- Category
jev-1.14.0- Pattern tags
jev-1.14.0
Evidence
These signals use repository metadata, an existing summary and at most 700 cleaned README characters. ≥ 0.60 means reported in that excerpt; ≤ 0.40 means not found there. The middle band is inconclusive. None is a source-code audit or a test result. The runnable-tool question excludes research experiments; the recommendation question asks for a learning shortlist, not a quality verdict.
- Jev is core to the projectFound in material0.81
- Shows a System One patternFound in material0.85
- Describes handling uncertaintyFound in material0.78
- Reports measured numbersunclear0.45
- Is a runnable toolFound in material0.87
- Top-10 learning shortlist signalFound in material0.62
- Is a replica of the modelFound in material0.78
- Solves a common problem (0–2)score on a 0–2 scale1.12 / 2
- About JevFound in material1.00
no-judgment band (0.40–0.60)
no-judgment band (0.40–0.60)
The Jev model and date for each judgment stage are listed above. Card written by DeepSeek V4.1 Flash from the README on 31 Oct 2026.