goodwatch-monorepo
GoodWatch movie and TV recommendation monorepo, listed for an experiment comparing Jev question designs for attribute scoring.
Overview
GoodWatch is a live movie and TV discovery service offering personalised recommendations, a fingerprint that classifies titles by hundreds of categories such as emotions, plot, subgenres and visual style, streaming availability, and aggregated ratings. The monorepo contains the Remix web app, admin tools, Windmill data pipelines, Qdrant similarity search, CrateDB, MongoDB, Redis, monitoring and deployment scripts.
According to the list summary, the repository includes an experiment comparing Jev question designs and batch sizes for scoring film and TV attributes, which would map naturally onto the fingerprint categories. The README does not describe this experiment, so its results are not documented here. Development is coordinated through a living roadmap issue and a Discord community, and the README notes that its local setup guide is outdated.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Score | How strongly does this title exhibit a given attribute (e.g. an emotion or visual style)?(inferred) | — |
How it works
- Select titles and attribute categories
- Ask Jev with different question designs and batch sizes
- Compare results against existing classifications
Strengths
- Real-world attribute scoring use case at scale
- Attribute categories map naturally to typed questions
Caveats
- Jev experiment not described in the README
- Local run guide marked outdated
Getting started
See the repository and Discord; the README notes the local setup guide is outdated.At a glance
- Type
- Project
- Status
- research
- Language
- TypeScript
- Runtime
- Remix web app + data pipelines
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
A film-and-TV attribute-scoring experiment inside GoodWatch comparing Jev question designs and batch sizes.
logicrw/awesome-jev-projects
Frequently asked questions
What is GoodWatch?
A movie and TV recommendation and discovery service.
How is Jev used?
In an experiment comparing question designs and batch sizes for attribute scoring, per the listing.
Are results published?
Not in the README.
Related projects
- jev-arena★ 122
Side-by-side tool comparing Jev and DeepSeek on labelling the same comments for relevance, sentiment and intent, with replay and reports.
ProjectbetaTypeScript Bun harness that runs Jev over datasets via OpenRouter, streams JSONL answers, and checks calibration, stability and regressions.
LibrarybetaTypeScriptLive race labelling 1,000 app reviews with four typed Jev questions versus Gemini 3.8 Flash JSON output.
AppprototypeJavaScriptClassifies 1,000 arXiv AI papers into 24 topics with one Jev Choice each and checks agreement against an LLM judge.
ProjectprototypePythonProfile matching that encodes preference criteria as Clef or Jev decision vectors, retrieves with HNSW and checks rules both ways.
BenchmarkresearchPython- jev-samples★ 1
Runnable Jev samples: README checks, AGENTS.md readiness, PR triage, MovieLens next-watch ranking and issue triage, with measured runs.
ProjectbetaPython