jev-py-integration
Support-ticket triage pipeline where Jev scores every ticket and Python decides whether a general LLM should write a reply.
Overview
This folder-sized project is a small support-ticket triage pipeline that illustrates System One versus System Two. Every incoming ticket is judged by Jev in a single call with several typed questions: Nouls for whether it is urgent, whether the customer is angry and whether it needs a human, and a Choice for intent (refund, technical help, spam or other). Plain Python if statements then decide whether the ticket is worth a second pass by a general LLM from OpenAI or Anthropic, which writes the reply.
The README is largely explanatory, with diagrams contrasting parsing an LLM's free-text answer with reading a typed probability and comparing it to a threshold. It walks through one ticket step by step, includes a file guide and run instructions, and suggests tuning and next steps. The point it makes is that the LLM is only called when Jev's scores say it is worth it.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | is_urgent: is the message time sensitive? | yes · no |
| Yes / no | needs_a_human: does this ticket need a human? | yes · no |
| Choice | intent: what does the customer want? | refund · technical_help · spam · other |
How it works
- Send each ticket to Jev with Noul and Choice questions
- Compare probabilities with thresholds in Python
- Call an LLM to draft a reply only for flagged tickets
- Route or reply accordingly
Strengths
- LLM costs only for tickets that need it
- Clear educational write-up of the pattern
- All judgments in one round trip
Caveats
- Small example rather than a product
- Needs TypeSafe and an LLM provider key
- Thresholds need tuning
Getting started
See the Running it section of the README.At a glance
- Type
- Project
- Status
- prototype
- Language
- Python
- Runtime
- Script
- Has code
- yes
Use cases and integrations
How curated lists describe it
Jev System one AI model, very fast decision making engine. Jev AI will use input tokens but no output tokens
logicrw/awesome-jev-projectsJev System one AI model, very fast decision making engine. Jev AI will use input tokens but no output tokens.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
Urgency, anger, need for a human and the ticket's intent.
When is the LLM called?
Only when Jev's scores cross thresholds set in Python.
Is Score used?
Not yet; the example uses Noul and Choice.
Related projects
Local-first routing engine for long-running agent tasks that uses a Jev decider to pick a model group per turn.
PluginproductionRust, TypeScriptOpen-source AI gateway for 100+ LLMs that can use Jev to classify requests for complexity routing and to compact tool results.
ProxyproductionPythonOpen-source workspace for running and reviewing AI coding agents across devices, with an optional Jev-based automatic model router.
ProjectproductionTypeScript- jev-router★ 593
Per-turn model routing for Claude Code and Codex: wraps the real CLI and lets Jev send simple turns to fast models and hard ones to strong.
ProxybetaJavaScript - jev-codex-router★ 274
Per-call router for Codex where Jev picks the model and thinking effort for every model call, including tool continuations.
ProjectbetaPython Open-source team agent platform for Slack, GitHub and more that asks Jev once per session which agent runtime and model to use.
ServicebetaTypeScript