Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

jev-minimis-watch

French toolkit that prepares de minimis state-aid cumulation reviews, with Jev classifying sourced cases into review categories.

ModelprototypeMIT
View on GitHub ↗

Overview

Jev Minimis Watch turns a sourced cumulation of de minimis state aid into an explicit, reviewable category. Documentation is in French. The repository separates rules that can be checked in code (identifiers, dates, exact amounts, filters and clear-cut cases) from the semantic comparison given to Jev, whose question and criteria are versioned in the source.

Jev chooses between clear_margin, review_required, likely_exceeded and no_effect. Low confidence (for example 0.62 in the bundled scenario) forces a human review flag so no automatic action is triggered, and a domain edge case resolves deterministically with zero Jev calls. The output is meant to order or prepare human review and is explicitly not an administrative decision, legal advice or eligibility guarantee. Demos use only synthetic data and a fake provider, so they make no network calls and do not measure Jev's real quality. It references the French public de minimis aid register.

What the decision model decides

TypeQuestionAnswers
ChoiceHow does this aid cumulation stand against the de minimis ceiling?clear_margin · review_required · likely_exceeded · no_effect

How it works

  1. Collect a sourced aid record with amounts, dates and linked entities
  2. Apply deterministic rules in code; settle clear cases without Jev
  3. Ask Jev to classify remaining cases
  4. Flag low-confidence results for human review
  5. Output a label to prioritise the review queue

Strengths

  • Clear split between code rules and model judgment
  • Uncertainty stays visible through a review flag
  • Tested with CI and a fake provider

Caveats

  • Demos use synthetic data only and do not measure real accuracy
  • Not legal advice or an administrative decision
  • French-language documentation

Getting started

git clone https://github.com/gbesse/jev-minimis-watch.git && cd jev-minimis-watch && npm install && npm run demo

At a glance

Type
Model
Status
prototype
Language
JavaScript
Runtime
Node library
Has code
yes
Has tests
yes
License
MIT

Use cases and integrations

How curated lists describe it

Frequently asked questions

What does Jev decide here?

Whether a sourced aid cumulation has a clear margin, requires review, has likely exceeded the ceiling, or has no effect.

Is the output a legal decision?

No. It only prepares and orders human review.

Do the demos call Jev?

No. They use a fake provider and synthetic data, making no network calls.

Related projects

  • jev-workflow

    Toolkit for Jev decision contracts: compiled policies, egress controls, tracing, stability checks and certified decision SLOs.

    ProjectbetaTypeScript
  • jev-pii

    Local-first personal-data inventory tool with optional Jev classification per column, not per row, and dry-run payload review.

    ProjectprototypePython
  • jev-norma

    Spanish compliance triage: clauses extracted deterministically, then Jev labels each obligation, risk or no action.

    LibrarybetaPython
  • Jev Policy Engine

    Python policy-as-code SDK that runs Noul, Choice and Score rules defined in YAML through Jev, with audit mode and fail-closed options.

    LibraryprototypePython
  • jev-uipath-coded-agent

    UiPath coded agent for AML alert triage where an LLM reads and writes but Jev makes every red-flag, risk and disposition decision.

    AppprototypePython
  • TypeSafe AI Playground

    Rust CLI of Jev experiments: HIPAA-style PHI detection, code-comment review, tone analysis and occupation classification.

    CLIprototypeRust