jev-boe-demo
Daily pipeline that reads Spain's official gazette with Jev, scoring relevance, tagging categories and picking a summary paragraph.
Overview
BOE para ciudadanos is a Spanish-language demo that applies Jev to Spain's official state gazette (BOE). Each morning a pipeline downloads the day's sections I, II-B and III and asks Jev, in one call per provision, a set of closed questions: how much does this change things for ordinary people (a 0-3 Score, relevant if 2 or more), whether it contains a concrete measure about each of 13 categories such as housing, employment or taxes (Nouls assigned above 0.6), which paragraph best explains what changes and for whom (a Choice over the paragraphs), and for public-job announcements whether the call is new and open.
Because Jev does not generate text, each summary is an original BOE paragraph selected by the model rather than written by it. The README reports a full day of 50-80 provisions (about 300k tokens) analysed in 5-15 s with six parallel calls for about $0.01, recording real latency per call. The site is a static Astro build on GitHub Pages with daily JSON data in the repository. The author notes the code was written with Claude Code.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Score | How much does this change things for ordinary people? | 0 · 1 · 2 · 3 |
| Yes / no | Does this contain a concrete measure about housing / employment / taxes / ...? | yes · no |
| Choice | Which paragraph best explains what changes and for whom? | — |
How it works
- Download the day's BOE sections
- One Jev call per provision with all questions
- Apply thresholds for relevance and categories
- Store JSON and publish the static site
Strengths
- Summaries are verbatim official text, not generated
- Full day processed in seconds for about $0.01
- Static site with no login
Caveats
- Spanish only
- Category thresholds are fixed by the author
Getting started
`pnpm install`, copy .env.example to .env with TYPESAFE_API_KEY, then `pnpm pipeline --today`.At a glance
- Type
- App
- Status
- prototype
- Language
- TypeScript
- Runtime
- Node pipeline, Astro static site
- Has code
- yes
Use cases and integrations
How curated lists describe it
Daily demo applying TypeSafe's Jev model to Spain's official gazette (BOE).
wh000wh000/awesome-jev-liveDaily demo applying TypeSafe's Jev model to Spain's official gazette (BOE).
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
Relevance to ordinary people, topic categories and which paragraph best summarises each provision.
Are the summaries generated?
No, they are original BOE paragraphs chosen by Jev.
How fast is it?
A full day's gazette in 5-15 s for about $0.01, per the README.
Related projects
Chinese-language Zotero AI reading assistant that uses Jev judgments to classify papers in your library.
LibraryproductionJavaScriptClassifies tax PDF pages into 261 IRS forms and 7 page kinds with one Jev request per page and a confidence gate.
ProjectproductionTypeScriptRust tool that asks Jev which PDF pages genuinely need OCR, extracting text pages locally and sending only the rest to OCR.
ProjectbetaRustMulti-user Telegram bot that judges every new post in public channels with Jev templates and routes matches to your DM or channels.
ProjectbetaPythonDocker service that classifies the Paperless-ngx inbox with Jev (or local Clef via Ollama), setting confident fields and queueing the rest.
ProjectbetaLibrary, CLI and app that classify PDFs, DOCX and PPTX with Jev and split document packets at predicted boundaries.
BenchmarkbetaPython