Why NotJev : Kev
Chat models are great at writing. They are awkward as product gates. NotJev : Kev turns an LLM into a typed classifier your code can threshold.
vs chat LLMs
| You need | Chat gives you | Kev gives you |
|---|---|---|
| A label | Prose you must regex | A typed key + probabilities |
| A gate | “I think so…” | noul you can threshold |
| Confidence | Vibes | Concentration of the distribution |
| Many judgments | N serial prompts | One request, questions in parallel |
| Control | Vendor lock-in | Your GPU / API / laptop |
vs hosted Jev / OpenJev
| Hosted Jev / OpenJev | NotJev : Kev | |
|---|---|---|
| License | Proprietary / mixed | Apache-2.0 |
| Deploy | Their cloud / weights | Self-host anywhere |
| Models | Fixed stack | BYO — Ollama, vLLM, OpenAI, … |
| Wire format | System One | Same shape (choice / score / noul) |
| Offline | No | Mock backend for CI & demos |
| DX | API key | Playground · SDK · CLI · MCP |
| Training | N/A for you | None — inference + API only |
Benchmark edge
On the public OpenJev held-out suite with qwen3.5:9b, NotJev : Kev reaches 83% Banking77 (vs ~82% published Jev held-out / 80.3% JevBench), plus strong CLINC, AG News, SST-5 within-1, and toxicity noul numbers — without locking you into a closed model.
Independent project. Not affiliated with TypeSafe AI or OpenJev.