Does the triage demo go beyond printing probabilities?

Yes, that is the point: one demo maps failures to Choice root cause, Score release impact, and Noul retry guidance, then routes each to filing, fixing, pinging DevOps, retrying, or human review.

Why do the Gateway demos fail with verification or auth errors?

Add a card to the Vercel team for verification, check the gateway key in .env, and re-run after a wait on throttling; the repo documents each failure shape.

What do the guardrail demos actually assert about agents?

One guards whether a test agent should run a command and another checks whether a RAG answer is grounded, giving a small LLM-as-judge replacement pattern for test pipelines.

How does it teach the documented Jev weak spots safely?

Run the breaker script that probes math, dates, counting, double negation, and injection weak spots, then shows the fix; differing results are treated as real model behavior.

What does the triage demo actually route?

Eight failed CI tests go in, and Jev answers root cause (Choice), release impact (Score) and whether a retry will pass (Noul). Code then routes each test to file bug, fix test, ping DevOps, auto-retry or human review, rather than just printing probabilities.

Can I run the demos without any cloud key?

Yes. npm run demo:local runs the same triage workflow through Ollama's /v1/systemone endpoint with no API key or cloud service, and a Postman collection is included. The cloud demos need AI_GATEWAY_API_KEY and can fail with customer_verification_required if the Vercel team has no card on file.

What does the break demo probe, and is a failure bad?

It targets Jev's documented weak spots, including math, dates, counting, double negation and prompt injection, then shows the code-side fix. A FAIL is expected and the README treats it as good demonstration material rather than a bug.

CommunityJev Integrations & Adapters

Jev Qa Demos

Video companion code showing Jev for QA work through the Vercel AI Gateway, from CI failure triage to guardrails, grounding checks, and local Ollama runs.

See highlights

Open github.com

Stars: 3
Language: TypeScript
Published by: Community
Author: gauravkhuraana
Last verified: Oct 2026

gauravkhuraana/jev-qa-demos

User repository on GitHub

View on GitHub

Jev (TypeSafe AI) demos for QA / SDET engineers via Vercel AI Gateway - simple, commented TypeScript for a video walkthrough

Forks: 3
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

Demos for the video "Jev explained for testers": TypeSafe AI's System One model, called through Vercel AI Gateway with the AI SDK's experimentalevaluate. Your Vercel team needs a card on file, even for free usage. Otherwise calls fail with customerverificationrequired. Run everything in video order with one command: TypeSafe's API calls the yes/no primitive a Noul (from Bernoulli). Vercel's AI SDK calls the same thing type: 'boolean', with the answer in .probability. 401 / 403: check AIGATEWAYAPIKEY in .env.

Quoted from the project's README. Read the full README.

Tagscommunity github

Works forJev for Classification

Used inJev for Software & DevTools

Languages: TypeScript

Highlights

Smallest-call smoke test prints one probability so gateway wiring can be checked in minutes.
CI triage demo routes known failures with Choice, Score, and Noul questions plus table-driven unit tests.
Guardrail and grounding demos cover command gating and answer checking, with a local Ollama variant.
Includes a breaker script for weak spots, a Postman collection, and troubleshooting for gateway errors.

Watch out

Demo scripts call a live gateway so behavior can vary run to run, and free-tier access may need verification on the Vercel team. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

n8n-nodes-jev

Community n8n node that adds Jev Choice, Score, and Noul questions to workflows, keeping native answers, model version, token usage, and latency for confidence routing.

Jev Integrations & Adapters#community#github

Community

rahulthakore16/n8n-nodes-jevTypeScript

omp-plugin-jev-router

Oh My Pi extension that asks Jev whether each interactive prompt is simple or advanced, then switches the OMP model role and thinking level without rewriting the prompt.

Jev Integrations & Adapters#community#github

Community

atulify/omp-plugin-jev-routerTypeScript

SiYuan

46.7k

Privacy-first, self-hosted personal knowledge management system with block-level references and a local-first Markdown editor; the project integrates Jev's decision model for high-speed document classification, flashcard intent categorization and automated tag routing.

Jev Integrations & Adapters#community#note-taking#knowledge-base

Community

siyuan-note/siyuanTypeScript

Eliza

19.6k

Open-source TypeScript agentic operating system: a core runtime, app for web and mobile, CLI, cloud services and first-party plugins. The directory's curated summary reports it integrates TypeSafe System One decision services for sub-100ms intent classification, action dispatching and confidence-gated tool execution.

Jev Integrations & Adapters#community#agents#framework

Community

elizaOS/elizaTypeScript

Inbox Zero

12.4k

Open-source AI email assistant (Next.js, Postgres, 12k+ stars) with a decision-model layer that can run on openrouter:typesafe/jev-1.13: Jev-backed questions cover rule selection, cold-email detection, thread status, recurring patterns, unsubscribe pages, and sender categorization, each with evals.

Jev Integrations & Adapters#community#email#gmail

Community