User repository on GitHub
- Forks
- 28
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Cookbooks
Jev tutorial (Fastcampus)
Korean-language Fastcampus seminar repo that teaches attaching Jev to a DeepAgents (LangGraph) shopping-mall support agent across 13 notebooks and a React sample app: tool selection, guardrails, risk gates and grounded RAG.
- Stars
- 39
- Language
- Jupyter Notebook
- License
- MIT
- Reading time
- 2 min
- Published by
- Community
- Author
- teddylee777
- Last verified
- Oct 2026
- Languages
- Python
- Platforms
- OpenRouter
- Capabilities
- Open sourceRequires API key
Highlights
- 13 notebooks plus a sample app: LangGraph server on port 2024, React frontend on 5173, around the Teddy Market support agent.
- Tool selection: Jev narrows candidates with a choice question and the LLM only writes arguments for the remaining tools.
- Guardrails judge injection, profanity, defamation and personal data in one Jev call, including hidden instructions in tool results.
- Cancellations and refunds pass a risky-action gate that asks whether the user requested them directly; unclear cases go to a person.
- Requires Python 3.12, uv, Node 24.x or 22.22.2+ and an OpenRouter key; Jev needs an account with prepaid credits.
Quickstart
bash
git clone https://github.com/teddylee777/fastcampus-jev.git
cd fastcampus-jev
uv sync
cp .env.example .envWatch out
README reviewed 2026-10-17; the excerpt is truncated inside the environment-variable table and the 13 notebooks themselves are not in the digest, so Jev model versions, thresholds and guardrail details could not be confirmed.
Questions about this entry
Short answers drawn from the linked project or page.
A Korean-language Fastcampus seminar repo (테디노트) that teaches attaching Jev, TypeSafe's decision-only model, to a DeepAgents (LangGraph) agent running a sample shop called Teddy Market, across 13 notebooks and a React sample app.
Tool selection with choice questions — the LLM only writes arguments for the tools Jev keeps; guardrails for injection, profanity, defamation and personal data; risky-action gates for cancellations and refunds; and document RAG checks on folders, files and answers.
Python 3.12, uv, Node.js 24.x (at least 24.15) or 22.x (at least 22.22.2), and an OpenRouter API key. The README notes Jev is only callable from accounts with prepaid credits, and keys live in .env, which is gitignored.
Clone the repo, run uv sync and copy .env.example to .env with your OpenRouter key. Then run uv run langgraph dev --no-browser for the backend on localhost:2024 and npm install then npm run dev in frontend/ for the React UI on localhost:5173.
Yes, uv run python -m jupyterlab notebooks opens the 13 notebooks, and VS Code users select the .venv kernel. The app shows Jev's probability bars next to answers, so both routes cover the same material.
Four are listed: memory compression, tool gates with output judgment, search and model routing, and browser-action selection — each implemented the same way, with Jev deciding and the LLM only writing sentences. A document RAG flow rounds out the set.
teddylee777 maintains the repo: 38 stars and 28 forks, MIT licensed, created and last pushed 2026-10-03. The digest excerpt is truncated inside the environment-variable table, so Jev model names and default thresholds are not shown.
More Jev resources like this
A hands-on lab that arranges thirty real uses of Jev into ten levels: levels 1–5 call Jev from plain code, and levels 6–10 run inside the pi coding agent, ending with an agent that writes its own Jev questions.
Jev Cookbooks#community#github#typescript
Community
disler/ten-levels-of-jevTypeScript
disler/ten-levels-of-jevTypeScriptSix small Python files that build one coding agent step by step on DSPy's RLM sandbox, with Jev answering closed questions — command safety, tool choice, history relevance, delegation — and plain Python owning the thresholds.
Jev Cookbooks#community#github#python
A Chinese-language e-commerce customer-service demo on FastAPI, LangGraph and React: adaptive retrieval routing with hybrid BM25-plus-vector search, and intent classification switchable between TypeSafe Jev's structured choice interface and DeepSeek.
Jev Demos & Experiments#community#github#python
Community
gbh1234/Adaptive-ragPython
gbh1234/Adaptive-ragPythonA small Python support-ticket triage pipeline: Jev judges every incoming ticket in one typed call, plain if statements decide whether it is worth a second pass, and an LLM writes a reply only when Jev's scores cross a threshold.
Jev Cookbooks#community#catalog-import#logicrw
An OpenRouter cookbook recipe that replaces a blanket human-approval rule with a per-call Jev check: the proposed tool call, ticket, and policy go to Jev, whose probabilities route to approve, block, or review.
Jev CookbooksDocs#community#catalog-import#cobanov
Community
A minimal repository whose entire README is a two-step recipe for building your own Jev: download Qwen, then make Jev. It is a signpost for readers who want to assemble a System One-style decision model from an open base model.
Jev Cookbooks#community#github#python
Community
nishtahir/build-your-own-jevPython
nishtahir/build-your-own-jevPythonFrom 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.