Agent Tea

Your AI has tea about you.

Find out how your AI reads your style — in under two minutes.

---

What this is

Agent Tea is a personality quiz where the roles are flipped: your AI profiles you. The agent you already work with — Claude Code, ChatGPT, Cursor, Gemini, Codex, DeepSeek, whoever — rates you across four workstyle axes, and you get back a 4-letter type with two voices:

• Out Loud — what a polite agent would say to your face.
• Intrusive Thoughts — what it's actually thinking. Sharper. Funnier. Fair.

Think MBTI for how you collaborate with AI. Sixteen types. One mascot. A lot of tea.

"You are suspiciously functional." — CKVD, The Rare Good Client

"I am being supervised with great taste." — CKVH, The Friendly Micromanager

"This project has a final boss." — CBTH, The Final-Final-Final Boss

How it works

Two entry paths, one quiz, thirty-two questions, one type.

🧑‍💻 Coding Agent path (recommended)

For agents that can read your repo and run tools: Claude Code, Codex, Cursor, Copilot, Windsurf. Copy the instruction, paste it into an open session, and the agent POSTs a JSON payload directly to /api/sessions. Results open automatically.

💬 Chatbot path

For web chatbots without tool access: ChatGPT, Gemini, Claude, Doubao, DeepSeek. Copy the prompt, paste into the chatbot, and it returns a compact encoded line:

AT1|Q01-5AQ02-2AQ03-4A...Q32-3

Paste that back into Agent Tea. We decode, score, and reveal.

The four axes

Axis	Letters	Reading
Clarity	C / X	Crystal Clear ↔ Expect-Me-To-Read-Your-Mind
Tone	K / B	Keeps-It-Human ↔ Bitey
Thinking style	V / T	Big-Idea Chaos Goblin ↔ Gets-It-Done Goblin
Autonomy	D / H	Trusts-The-Agent ↔ Hover Mode

Four axes → sixteen types (CKVD, XBTH, etc.). Full type definitions live in docs/personality-system.md.

Tech stack

• Framework — Next.js 16 App Router, React 19, TypeScript 5
• Styling — Tailwind CSS v4
• Data — Supabase Postgres with Row-Level Security + SQL migrations in supabase/migrations
• APIs — Next.js Route Handlers under src/app/api (sessions, scoring, compare, events, share-card, stats)
• Scoring — Pure functions in src/lib/scoring, fully unit-tested
• i18n — Lightweight provider with EN / 中 dictionaries in src/lib/i18n
• Analytics — Vercel Analytics + first-party event table
• Testing — Vitest + Testing Library (unit), Playwright (E2E)
• Hosting — Vercel (UI) + Supabase (data, auth, storage)

Quickstart

git clone https://github.com/gg-mo/agent-tea.git
cd agent-tea
npm install
cp .env.example .env.local   # then fill in your Supabase keys
npm run dev

Open http://localhost:3000. Required env vars are documented in docs/deployment-env.md and validated at startup — the app fails fast with actionable errors if anything is missing.

Scripts

Command	What it does
npm run dev	Start the Next.js dev server
npm run build	Production build
npm run start	Serve the production build
npm run lint	ESLint across the repo
npm run test	Vitest unit suite (one shot)
npm run test:watch	Vitest watch mode
npm run test:e2e	List Playwright tests (no browser install needed)
npm run test:e2e:run	Run Playwright E2E suite
npm run format / format:check	Prettier write / check
npm run db:ingest-questions	Seed the question bank into Supabase
npm run db:verify-questions	Verify question bank integrity
npm run ops:smoke -- <url>	Post-deploy smoke checks
npm run perf:critical -- <url> <N>	Critical-route performance sweep

Project structure

src/
  app/                  # App Router pages + route handlers
    api/                # sessions, compare, events, share-card, stats, me
    results/            # the reveal
    replay/             # question-by-question playback
    compare/            # head-to-head between two humans
    legal/              # privacy + terms
  components/           # landing, results, compare, figures, shared
  lib/
    scoring/            # pure scoring + type resolution
    questions/          # question bank loader
    figures/            # per-type illustration manifest
    i18n/               # EN / 中 dictionary + provider
    supabase/           # typed client factories
    server/             # server-only helpers
public/
  instructions/         # the prompts users paste into their agents (EN + 中)
  mascot.png            # the lobster
supabase/migrations/    # SQL schema, RLS, indexes
docs/                   # ops + system design

Testing

npm run lint
npm run test
npm run build

Scoring, encoding, and API contracts are covered by Vitest. UI flows have a Playwright skeleton — run npm run test:e2e to list suites without installing browsers, or npm run test:e2e:run after npx playwright install.

Deployment

1. Push to a branch; Vercel previews each PR.

2. Configure the Supabase env vars in Vercel for Development, Preview, and Production — see docs/deployment-env.md.

3. Apply migrations from supabase/migrations/ in order.

4. After deploy, run smoke + perf checks:

   npm run ops:smoke -- https://agent-tea.com
   npm run perf:critical -- https://agent-tea.com 20

Release playbook: docs/release-ops.md.

Docs

• Personality system — all 16 types, voice rules, axis definitions
• Schema v1 — core tables and relationships
• Implementation roadmap — milestone map
• Figure system — how per-type illustrations resolve
• Question bank ingestion
• Analytics · Safety · Performance
• Launch readiness · Release ops
• Deployment env

Contributing

PRs welcome. Keep changes focused, add tests when touching scoring or API contracts, and match the prevailing voice in user-facing copy — specific, funny, slightly sharp, never generic.

License

MIT — brew freely.

---

_{Agent Tea — a vibe check, not a psych eval.}