MeetingMind — AI Meeting Summarizer

A full-stack AI-powered meeting summarizer that turns raw transcripts into structured summaries with action items, key decisions, and follow-up questions.

Stack: Next.js 16 + Tailwind CSS + Shadcn/ui + Anthropic SDK (Claude) + Supabase

Features

• Paste or upload meeting transcripts (.txt, .md)
• AI-generated summaries with structured sections: Summary, Key Decisions, Action Items, Follow-ups
• Real-time streaming — see the summary generate token by token
• History sidebar — past summaries saved in Supabase (optional)
• Export — copy to clipboard or download as .md
• Sample transcripts — try it instantly with built-in demo data
• Dark mode UI

Setup

1. Install dependencies

cd meeting-summarizer
npm install

2. Configure environment variables

Edit .env.local:

ANTHROPIC_API_KEY=your-anthropic-api-key

# Optional — Supabase for saving history
NEXT_PUBLIC_SUPABASE_URL=your-supabase-url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your-supabase-anon-key

3. (Optional) Set up Supabase

Create a free project at supabase.com, then run this SQL in the SQL Editor:

CREATE TABLE summaries (
  id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
  title TEXT NOT NULL,
  transcript_preview TEXT,
  summary_markdown TEXT NOT NULL,
  action_items JSONB,
  created_at TIMESTAMPTZ DEFAULT now()
);

ALTER TABLE summaries ENABLE ROW LEVEL SECURITY;
CREATE POLICY "Allow anonymous read" ON summaries FOR SELECT USING (true);
CREATE POLICY "Allow anonymous insert" ON summaries FOR INSERT WITH CHECK (true);

The app works without Supabase — you just won't have the history feature.

4. Run the app

npm run dev

Open http://localhost:3000.

API

POST /api/summarize

Streams an AI-generated meeting summary.

Request body:

{ "transcript": "Sarah: Let's start the meeting..." }

Response: Server-Sent Events stream with chunks:

data: {"type": "chunk", "content": "## Summary\n"}
data: {"type": "chunk", "content": "The team discussed..."}
data: {"type": "done"}

Tech Stack

Layer	Technology
Framework	Next.js 16 (App Router)
Language	TypeScript
Styling	Tailwind CSS + Shadcn/ui
AI	Anthropic SDK (Claude claude-opus-4-6)
Database	Supabase (PostgreSQL)
Deployment	Vercel (recommended)

Customization

• System prompt: Edit lib/prompts.ts to change summary format or style
• Sample transcripts: Add your own in lib/prompts.ts → SAMPLE_TRANSCRIPTS
• Model: Change MODEL in lib/prompts.ts (default: claude-opus-4-6)
• Theme: Toggle dark/light in app/layout.tsx (remove dark class for light mode)

Deploy to Vercel

1. Push to GitHub
2. Import in Vercel
3. Add environment variables in Vercel dashboard
4. Deploy