Autonomous Memory for the Internet

Production-ready Next.js frontend + API layer for corporate claim accountability.

Stack

• Next.js 14 + TypeScript + App Router
• Tailwind + shadcn-style components
• React Query
• Recharts + react-force-graph-2d
• PostgreSQL (required)
• OpenAI (required for /api/scan)
• Modulate (required for /api/voice)

Fast Start (Real Data)

1. Install deps

npm install

2. Configure env

cp .env.example .env.local

Fill these values:

• DATABASE_URL
• OPENAI_API_KEY (for real /api/scan extraction)
• MODULATE_API_KEY
• MODULATE_BASE_URL
• TAVILY_API_KEY (for live /api/search and /api/outbreak)

3. Create DB schema

psql "$DATABASE_URL" -f db/schema.sql

4. Start app

npm run dev

API Endpoints

• GET /api/search?q=
• GET /api/entity/:id
• GET /api/entity/:id/claims
• GET /api/entity/:id/graph
• GET /api/entity/:id/voice
• GET /api/claim/:id/evidence
• GET /api/outbreak
• GET /api/report/:id
• POST /api/scan (OpenAI extraction)
• POST /api/voice (Modulate integration)

POST /api/scan body

{
  "subjectEntityId": "comp-your-company",
  "speakerEntityId": "person-optional",
  "transcriptText": "full transcript or excerpt",
  "sourceUrl": "https://...",
  "sourceTitle": "Q2 Earnings Call"
}

Notes

• No client-side mock API usage remains.
• If env vars are missing, API returns 503 with clear error details.