QA Engine

AI-powered QA tools built with Next.js (App Router) and the official @anthropic-ai/sdk.

Live demo of two everyday QA workflows:

Tool	Input	Output
Test Case Generator	A feature described in free text	Structured test cases in tables — positive, negative, edge cases, plus open questions for the team. Each case has an ID, steps, and expected result.
Bug Report Formatter	A bug described in casual language	A professional report: title, justified severity, environment, steps to reproduce, expected vs actual result, and technical notes.

Architecture

• app/api/generate/route.ts — server-side API route that calls Claude (claude-haiku-4-5, chosen to keep demo costs low). The API key lives in the ANTHROPIC_API_KEY environment variable and never reaches the client.
• Streaming — the route streams text deltas back to the browser, so the result renders progressively as Claude writes it.
• app/page.tsx — single-page UI with a tab toggle between the two tools, a preloaded fictional example per tool, and Markdown rendering (react-markdown + remark-gfm) for the streamed tables.

⚠️ This is a demo tool — do not paste confidential or proprietary information.

Run locally

npm install
cp .env.example .env.local   # then put your real key in .env.local
npm run dev

Open http://localhost:3000.

Deploy to Vercel

1. Push this repository to GitHub.
2. In Vercel, click Add New → Project and import the repo. Vercel auto-detects Next.js — no build settings needed.
3. In Settings → Environment Variables, add:
   • ANTHROPIC_API_KEY = your Anthropic API key (get one at platform.claude.com)
4. Click Deploy.

Every push to the default branch triggers a new deployment.

Notes

• The API route runs on the Node.js runtime with maxDuration = 60 so longer generations aren't cut off (Vercel Hobby allows up to 60s for streaming functions).
• Input is capped at 4,000 characters and the response at 4,096 tokens to keep per-request cost predictable.

---

Built by Edgardo Morales Cosme