thirumurugan240/rag_with_jev_ai
User repository on GitHub
- Forks
- 1
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Demos & Experiments
RAG powered by Jev
A beginner-friendly RAG app that answers questions about your documents using only what is written there: rag.py finds the right chunks, asks Jev to check them and gets the answer from OpenAI, while a Streamlit page shows the result with Jev's scores.
- Stars
- 3
- Language
- Python
- Reading time
- 2 min
- Published by
- Community
- Author
- Thirumurugan240
- Last verified
- Oct 2026
- Languages
- Python
- Capabilities
- Open sourceRequires API key
Highlights
- Two small files: rag.py reads documents, finds the right chunks, asks Jev to check them and gets the answer from OpenAI.
- app.py is the Streamlit web page: a text box for your question and the answer shown with Jev's scores.
- sample_docs/ holds the documents the app knows about; drop your own .md or .pdf files in.
- README motivation: search returns 20 chunks but usually only 2 or 3 answer the question, and sending all 20 can confuse the model.
- GitHub metadata: no license, created and pushed 2026-10-06, one fork.
Watch out
README and GitHub metadata reviewed 2026-10-17; the excerpt stops inside the explanation of what Jev is, so setup instructions, scoring thresholds, model choices and any evaluation are not evidenced.
Questions about this entry
Short answers drawn from the linked project or page.
A beginner-friendly retrieval-augmented generation app: you ask a question about your documents and it answers using only what is written in them, per the README, with Jev checking the retrieved chunks before the answer is generated.
After retrieval, the app asks Jev to check the chunks found for the question, and the Streamlit page shows Jev's scores alongside the answer. The README frames Jev as the fix for search returning same-topic but wrong-answer chunks.
The sample_docs/ folder holds the documents the app knows about, and the README says to put your own .md or .pdf files there to make them part of the app's knowledge.
The README describes Jev from TypeSafe as the checker of retrieved chunks and OpenAI as the source of the final answer, so both need access. The truncated excerpt does not show the configuration or key setup.
A plain pipeline sends every retrieved chunk to the model. This one inserts a Jev checking step between retrieval and generation, and surfaces Jev's scores in the UI, which the README presents as protecting the answer from off-target chunks.
Thirumurugan240 publishes it with no license listed in the metadata; the repository was created and last pushed on 2026-10-06 and has 3 stars and one fork.
The excerpt stops inside the explanation of what Jev is, so install and setup commands, scoring thresholds, the OpenAI model used and any accuracy checks are not evidenced.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: A RAG loop where Jev decides what is real
X: Measuring what Jev does in a RAG pipeline
X: Jev-powered reranking for RAG. Likitd_’s jev-ranker uses Jev to rerank results in a RAG pipeline.
X: Coverage is not product value
More Jev resources like this
A small SeleniumBase project that sends a page's reduced HTML to Jev and gets back three structured answers: whether a CAPTCHA is present, a 0-100 confidence score and which of six types it is. The README warns scores vary between runs and are not calibrated probabilities.
Jev Demos & Experiments#community#github#captcha
Community
mdmintz/jev-testingPython
mdmintz/jev-testingPythonA Streamlit lab that places a bounded System One decision inside a coding-agent workflow: Laya classifies CI failures, LangGraph applies a fallback policy, and an OpenAI agent inspects a fixture repo and proposes a fix.
Jev Demos & Experiments#community#github#python
Community
natenberenstein/systemonePython
natenberenstein/systemonePythonA local deep-research tool that turns your own PDF, DOCX, PPTX, image, TXT, or Markdown files into a structured report, using TypeSafe's Jev to rerank Qdrant chunks and verify evidence coverage while an LLM plans and writes.
Jev Tools & CLIs#community#github#python
Community
oqura-ai/deepdocPython
oqura-ai/deepdocPythonA 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-ragPythonEleven small runnable projects for learning Jev and other System One decision models, each lesson pairing a one-command web demo with a whiteboard and README, from the three question types through jobs sorting and contract review to Jev inside a SQL query. Every demo makes live API calls.
Jev Demos & Experiments#community#github#python
Community
aianytime/jev-crash-coursePython
aianytime/jev-crash-coursePythonFive self-contained Jev demos, one folder each: Jev fetching for Claude, Jev and Laya grading 500 AI answers, a prompt-injection review gate, and two fine-tuned decider comparisons against Jev.
Jev Demos & Experiments#community#python#demos
Community
mani-aiml/jev-demosPython
mani-aiml/jev-demosPythonFrom 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.