User repository on GitHub
- Forks
- 1
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Tools & CLIs
Jev Route
A router that runs today on Jev's calibrated cloud decisions and turns your traffic into your own local routing model: every decision is logged with its full probability distribution, so the log is a labelled dataset you distill toward an air-gapped router.
- Stars
- 3
- Language
- Python
- License
- APACHE-2.0
- Reading time
- 2 min
- Published by
- Community
- Author
- mcftira
- Last verified
- Oct 2026
Used inJev for Software & DevTools
- Languages
- Python
- Capabilities
- Open sourceRuns locallyRequires API key
Highlights
- Routes on two axes at once - task difficulty and data sensitivity - plus a PII judgement and a task domain, with calibrated confidence as a routable fact.
- Every decision lands in decision-log/decisions.jsonl with its full probability distribution; the log is the training set for distillation.
- Graduation pipeline: export, train, evaluate against the teacher on held-out decisions, package, then switch backend.name from jev to distilled.
- v0.2's deterministic tier prefilter runs before any model call: when one tier survives, the backend call is skipped and logged with backend: prefilter.
- A two-layer local gate runs first: layer 1 deterministic and never bypassed, layer 2 a local semantic model in shadow by default; pinned-model bypass is opt-in.
Watch out
Grounded on the truncated README and GitHub metadata reviewed 2026-10-15; the v0.2 section is cut mid-sentence and the PyPI badge is commented out until the first release. Python 3.10-3.12 is targeted, and the distillation route expects enough of your own traffic to accumulate first.
Questions about this entry
Short answers drawn from the linked project or page.
A router for LLM traffic that starts on TypeSafe Jev's cloud decisions with one API key and zero training data, then turns your logged traffic into a local distilled model you can run air-gapped.
Jev answers the routing questions - difficulty, sensitivity, PII and domain - and every call is logged with soft probabilities. Once enough traffic accumulates, that log becomes the labelled set for distillation.
The documented pipeline is export, train, evaluate against the teacher on your own held-out decisions, package, then flip backend.name from jev to distilled. The README calls the cloud phase training wheels.
One Jev API key and zero training data, on Python 3.10, 3.11 or 3.12 per the badges. The README also lists mock, distilled, laya and shadow as decision backends in the routing chain.
It is deterministic and runs before any model call, ruling out tiers that could never serve the request. If exactly one tier survives, routing is free and the decision is logged with backend: prefilter.
Yes: decision-log/decisions.jsonl records the chain including soft probabilities, and prefilter decisions appear like any other so the distillation pipeline sees them.
The PyPI badge is commented out until the first release is published, so as reviewed it is a GitHub project at v0.2 under Apache-2.0; treat the distillation pipeline as an evolving product.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
Facebook: Intake Router keeps diagnosis out of the answer space at the front desk.
FacebookClinicnotes
X: OpenRouter Decision Model Rankings
X: herdr + jev, an agent router
X: A group chat bot with decision room routing
More Jev resources like this
Toolkit, knowledge base and local dashboard for building with Jev via OpenRouter Alpha Decisions or the TypeSafe Direct API: a zero-dependency SQLite request log, a localhost REST server and Python CLI demos.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
rileycarney/jevtoolsHTML
rileycarney/jevtoolsHTMLRoutes each prompt to a model tier with one Jev call: a Choice question classifies difficulty (simple, moderate, complex, frontier) and a Score rates stakes. Confidence-gated rules then fall back or escalate, and your application still makes the downstream request.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
gualican/jev-model-routerPython
gualican/jev-model-routerPythonTurns an inbox into a short action queue: Jev answers seven typed questions per email thread in one request, plain Python rules decide what enters the queue, and cached answers mean changing a rule never needs new API calls.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
tusharck/jev-inbox-queuePython
tusharck/jev-inbox-queuePythonA local, open-source personal AI assistant for your job: mail, chats and tickets become tasks, agents you already use such as Claude Code, Codex and Gemini do the work, and nothing is sent until you approve. The current release is v0.3.8.3.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
ldbumble/taskuaryPython
ldbumble/taskuaryPythonA Mac-only voice assistant: say Hey Jev or hold right Option and Jev decides every command in one call for $0.00004 per request, while Whisper transcribes locally, Fish Audio S2.1 Pro speaks replies, and an LLM wakes only for questions.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
henryklunaris/hey-jevPython
henryklunaris/hey-jevPythonA Rust CLI that streams Kubernetes logs, triages them with local offline rules or typed Jev questions online, and presents an interactive terminal UI plus JSON reports. Collection is read-only and every judgment is advisory.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
sunil-sadasivan/jevernetesGo
sunil-sadasivan/jevernetesGoFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.