jev-demo
Three Spanish-language demos where Jev decides and Claude reasons: Jira triage, code pre-flight and a PR review router.
Overview
jev-demo from LIDR Academy contains three demos on one pattern: state, typed question, Jev probability, your code decides, action. Demo 1 classifies Jira tickets by type, priority, team and whether a human is needed, then applies, applies and flags, or escalates. Demo 2 has Jev check each function before code is touched and Claude explains only the flagged ones. Demo 3 routes pull requests to rules with no LLM, to Haiku or to Sonnet, with Claude posting review comments on GitHub.
A cost script compares tokens and dollars for an LLM-only run (everything to Sonnet) against the Jev-assisted run. Everything calls real services; there is no simulated mode. It needs Node 22.18+ (TypeScript runs directly with no npm install), Claude Code logged in (called via claude -p), a TypeSafe key, a Jira Cloud token and the GitHub CLI. Documentation is in Spanish.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What type, priority and team does this Jira ticket have? | — |
| Yes / no | Does this ticket need a human? | — |
| Yes / no | Does this function already have problems? | — |
| Choice | Should this PR go to rules, Haiku or Sonnet? | rules · Haiku · Sonnet |
How it works
- Gather state from Jira, source code or a GitHub PR
- Ask Jev typed questions
- Code applies thresholds to act, flag or escalate
- Call Claude only where reasoning is needed and record costs
Strengths
- Real services end to end
- Cost comparison against an all-Sonnet baseline
- No npm install required
Caveats
- Needs several accounts (TypeSafe, Jira, GitHub, Claude Code)
- Spanish-language documentation
- Educational demo
Getting started
With Node 22.18+, Claude Code and TYPESAFE_API_KEY set, run the skills such as /triage, /pre-vuelo or /router from each demo folder.At a glance
- Type
- App
- Status
- prototype
- Language
- TypeScript
- Runtime
- Node scripts and Claude Code skills
- Has code
- yes
Use cases and integrations
How curated lists describe it
Jev decide, el LLM razona: triage de Jira, pre-vuelo de código y router de…
wh000wh000/awesome-jev-live
Frequently asked questions
What does Jev decide?
Jira ticket fields and escalation, which functions look problematic, and which reviewer tier a PR needs.
Is anything simulated?
No, every command calls the real services.
Do I need an Anthropic API key?
No, Claude is called through a logged-in Claude Code session.
Related projects
Glue that turns one Linux machine into a team of coding agents, with Jev making typed calls like who builds a task and whether it may merge.
ProjectbetaTypeScriptOpen-source team agent platform for Slack, GitHub and more that asks Jev once per session which agent runtime and model to use.
ServicebetaTypeScriptDesktop workspace for coding agents that runs tasks in parallel worktrees, listed as using Jev to pick the best agent per task.
ProjectbetaTypeScript, RustThree Italian-language Claude Code prompts that build Jev-powered model routing, context compaction and code review into your project.
ProjectprototypeProbot GitHub app that classifies, scores, checks completeness and dedupes issues and PRs with Jev, routing uncertain ones to humans.
ProjectbetaTypeScriptSmall Python framework and GitHub Action for Jev-powered workflows, with issue triage and PR screening built as examples.
ProjectbetaPython