jev-ai-use-cases
LangChain notebook of practical Jev examples: ticket triage, model routing, reply guardrails, tool selection and finance-inbox fraud checks.
Overview
Jev in Practice is a hands-on Jupyter notebook, runnable in Colab, with practical examples of Jev built on LangChain. Its framing is that an LLM writes and Jev decides: you give Jev the options and it picks one, reports how sure it is, and never answers outside the list.
Examples cover support-ticket triage (which team, how urgent, whether the customer wants a refund), model routing between small, premium and private models, guardrails on whether an AI-written reply is safe to send, tool selection with blocking of risky tool calls, and a finance inbox that classifies document type and urgency and flags bank-details-changed fraud. The notebook includes outputs from a real run. Jev is called through OpenRouter and the writing LLMs through Groq, and the README notes that langchain-typesafe is early alpha with experimental middleware.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which team should handle this ticket? | — |
| Choice | Small, premium or private model for this request? | small · premium · private |
| Yes / no | Is this AI-written reply safe to send? | yes · no |
| Yes / no | Does this email indicate bank details changed fraud? | yes · no |
How it works
- Install langchain-typesafe and set OpenRouter and Groq keys
- Open the notebook and run each example
- LLM drafts content; Jev makes the decision
- Inspect saved outputs from a real run
Strengths
- Five realistic, varied use cases in one place
- Runs in Colab with saved outputs
Caveats
- langchain-typesafe is early alpha
- Needs both OpenRouter and Groq keys
Getting started
`pip install "langchain-typesafe[experimental]" langchain-groq python-dotenv pandas`, add OPENROUTER_API_KEY and GROQ_API_KEY, then run jev_decision_guide.ipynb.At a glance
- Type
- Project
- Status
- prototype
- Language
- Python
- Runtime
- Jupyter / Colab (LangChain)
- Has code
- yes
- Has tests
- no
Use cases and integrations
How curated lists describe it
Hands-on LangChain examples of Jev, TypeSafe AI's decision model: support ticket triage, model routing, reply guardrails, tool selection and finance-inbox fraud checks. An LLM writes; Jev decides.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide in these examples?
Ticket routing and urgency, model choice, reply safety, tool selection and fraud flags.
How is Jev accessed?
Through OpenRouter via langchain-typesafe.
Is the LangChain integration stable?
No, the README says it is early alpha.
Related projects
Open-source team agent platform for Slack, GitHub and more that asks Jev once per session which agent runtime and model to use.
ServicebetaTypeScriptSupport-ticket triage pipeline where Jev scores every ticket and Python decides whether a general LLM should write a reply.
ProjectprototypePythonThree Spanish-language demos where Jev decides and Claude reasons: Jira triage, code pre-flight and a PR review router.
AppprototypeTypeScriptDependency-free Python wrapper that turns any LLM API into a Jev-style choose-one-option decision call for old machines.
ModelprototypePythonKorean course code for attaching Jev to Claude Code, Codex and OpenClaw: skills, hooks, routers and open-model comparisons.
ProjectproductionJavaScriptCompanion code for the ZazenCodes course on decision models: primitives notebook, use-case examples and a Pi model-router build.
ProjectproductionPython