User repository on GitHub
CV diagnostics and job alignment with TypeSafe Jev, React and FastAPI
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Tools & CLIs
Jev Resume Analyzer
CV review on localhost, built on TypeSafe System One (jev-latest): upload a PDF or DOCX, check the extracted text, and get whole-document judgments on work clarity, contribution, interpretability, outputs and scope, each expandable to the question, criteria, answer and probability distribution behind it.
- Stars
- 1
- Language
- Python
- Published by
- Community
- Author
- awun8191
- Last verified
- Oct 2026
About this project
Review a CV with fast, structured AI judgments, optionally against a job posting. Upload a PDF or DOCX, check the extracted text, and get a traceable report: every assessment shows the exact question, criteria, answer, and probability distribution behind it. Built on TypeSafe (System One model, jev-latest): structured judgments composed in code, not prompt-and-parse text generation.
Quoted from the project's README. Read the full README.
Works forJev for Classification
Used inJev for Software & DevTools
- Languages
- Python
Highlights
- Four outcomes per question - present, absent, not_applicable, unable_to_assess - each with a real distribution instead of one verdict.
- Explanations are predefined rubric text, clearly labelled, never generated advice; each finding expands to its question, criteria and per-option probabilities.
- One batched Jev request per analysis; failed requests never produce a fabricated report, and no CV text is stored in localStorage or a database.
- FastAPI backend with pypdf and defusedxml for DOCX traversal, React 18 and Vite frontend behind a Zod-validated API contract.
Watch out
No overall rating is shown yet: score maps stay null until they are calibrated on reviewed examples. Assessments are full-document only, with no fabricated bullet-level findings, and there are no hiring predictions, ATS scores, truthfulness verdicts or generated rewrites. Generated suggestions are deferred to a later phase, as are OCR, accounts and public deployment. Explanations are predefined rubric text, not generated advice, so findings read as rubric categories rather than tailored comments. It is a localhost demo with no accounts and no database, but CV text is sent to TypeSafe for analysis, so retention terms should be verified before any public deployment. Running it needs Python 3.11+, Node 20.19+ or 22.12+, uv and an API key, with the backend on port 8017 and the frontend on 5273. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
No. The design boundaries rule out hiring predictions, ATS scores, truthfulness verdicts and generated rewrites, and no overall rating is displayed yet because the scoring maps stay null until calibrated on reviewed examples. What you get instead are typed document-level judgments over work clarity, personal contribution, interpretability, outputs and results, scope, and optional summary specificity, plus two document-level questions when you paste a job posting.
No. Explanations are predefined rubric text, clearly labelled as such, so nothing in the report is generated advice. Every finding can be expanded to show the evaluated question, its criteria, the answer and the per-option probabilities. The rubric modules under backend/app/rubrics are what define what each question means, alongside the binding frontend-backend contract in docs/frontend-contract.md.
There are four answer outcomes - present, absent, not_applicable and unable_to_assess - each returned with a real probability distribution rather than a single value. A CV with no summary is not treated as a defect, and unable_to_assess is not collapsed into a zero, which is the point of keeping the distributions visible instead of scoring the absence of evidence as failure.
Python 3.11+, Node 20.19+ or 22.12+, uv and a TypeSafe key. The Vite dev server on 127.0.0.1:5273 proxies /api to a FastAPI backend on 127.0.0.1:8017, with an editable extraction preview between upload and analysis. There are no accounts, no database, no localStorage for CV text and no client-side keys, and a failed request never produces a report - but the text is sent to TypeSafe, so check retention terms before deploying this anywhere public.
In backend/ copy .env.example to .env, set TYPESAFE_API_KEY, run uv sync, then uv run uvicorn app.main:app --host 127.0.0.1 --port 8017; in frontend/ run npm ci and npm run dev, opening http://127.0.0.1:5273. Vite proxies /api to the backend.
Full-document assessments only, with no fabricated bullet-level findings; no overall rating yet, since scoring maps stay null until calibrated on reviewed examples; and no hiring predictions, ATS scores, truthfulness verdicts, or generated rewrites.
React 18, Vite, TypeScript, and a Zod-validated API contract on the frontend; FastAPI, Pydantic, pypdf, and defusedxml on the backend; and one batched TypeSafe System One request per analysis.
More Jev resources like this
39GitHub stars
Video tool where Jev scores subtitle segments and code assembles a short highlights cut plus a de-watered full cut with timestamped summaries.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
cclank/jevclipPython
cclank/jevclipPython30GitHub stars
Forkable weekday radar that asks Jev one yes-or-no question per interest over each new arXiv announcement and publishes a page plus RSS feed.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
Eliot5566/JEV-Paper-RadarPython
Eliot5566/JEV-Paper-RadarPython38GitHub stars
Async supervisory layer for Hermes Agent that locks Definitions of Done, proves mechanical criteria locally, and calls Jev only for unresolved semantic checks under a token budget.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
keeltrace/hermes-nervePython
keeltrace/hermes-nervePython19GitHub stars
Windows desktop aid that reads the local WeChat database for the foreground chat, strips identifiers, and asks Jev for stage, intent, tension, suggested moves, and a reply review.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
yushen100/wechat-jev-assistantPython
yushen100/wechat-jev-assistantPython48GitHub stars
Next.js dashboard plus Python feed that streams market data, computes indicators across timeframes, sends enabled states to Jev, and applies the judgments to a SQLite paper portfolio.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
zadescoxp/Jev-TradesPython
zadescoxp/Jev-TradesPython7GitHub stars
Tournament engine that ranks texts by pairwise or cohort duels judged with Jev or local open-weight models, using Elo and TrueSkill ratings with a live web dashboard.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
opaielsheikh/ai-elo-rankerPython
opaielsheikh/ai-elo-rankerPythonFrom the Jev guides
Original write-ups on what this entry is for.
The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.
Jev Tools & CLIs#gateway#vercel#cloudflare
Read article
Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.
Jev Tools & CLIs#open-models#local#onnx
Read article
Jev labels a record that already exists. Code moves, flags, or holds it. How the document, mail, and transfer tools keep the action out of the model.
Jev Tools & CLIs#classification#email#security
Read article
From the Jev community
Posts from builders shipping with Jev right now.