JevMeter
Score every sentence in a video, then render the results as a broadcast-style data overlay.
What Jev decides
For each configured dimension, whether the current sentence exhibits that quality.
JevMeter separates transcription and alignment from judgment, sends one multi-question Noul request per sentence, aggregates probabilities over time, and renders a 16:9 analytics video.
- Jev role
- Apply the same semantic yes-or-no dimensions to every spoken sentence.
- Input
- One aligned transcript sentence plus the configured judgment questions.
- Output
- Multiple Noul probabilities for that sentence.
- Decision frequency
- Once per transcribed sentence.
- Fallback
- Transcription, alignment, aggregation, and video rendering are deterministic pipeline stages and can be rerun separately.
How it works
Reported performance
Reported total Jev cost
0.0497 USD
Evaluate the two example sentence datasets documented in the README.
Author reported · Source reviewedCaptured Sep 21, 2026
Overview
JevMeter turns long-form video into a timeline of sentence-level judgments. Transcription and word alignment run first; each sentence is then evaluated on configured yes-or-no dimensions, and the probabilities are smoothed into charts, gauges, and overlays for a rendered output video.
Despite the visual meter, the current pipeline uses Noul probabilities rather than Jev Score. The judgments are not fact-checks, author evaluation is not independent, and users remain responsible for rights to downloaded source media.
- Language
- Python
- License
- MIT
- Author
- ChetasLua
Sources
Related projects
pg-jev
Use Jev decisions directly inside PostgreSQL filters, rankings, and classifications.
What Jev decides
Whether a row matches, how strongly it matches, or which supplied label best describes it.
jev-spam-eval
A reproducible email-classification study comparing zero-shot Jev decisions with trained TF-IDF baselines.
What Jev decides
Whether an email is ham, spam, or phishing in the main experiment.
Jev Search
Turn a plain-language request into source choices, search lanes, and visibly ranked links without generating an answer.
What Jev decides
Which query, sources, and time range fit the request, and how relevant each returned result is.