`README.md`

33 lines

# Trademarks & Copyright — Opportunity Finder

Searchable index of **abandoned US trademarks** and **public-domain copyrights** with a 4-agent SWOT analyzer,
composite scoring, news feed, and a draggable React Flow canvas for portfolio planning.

**⚠️ Not legal advice.** See `PLAN.md` for the disclaimer. Abandoned marks can carry common-law rights; "PD" works can still have derivative-work copyrights.

## Run

```bash
cp .env.example .env.local # add ANTHROPIC_API_KEY for SWOT
npm install
npm run db:init # creates schema in trademarks_copyright
npm run db:seed # ~30 real seed items
npm run dev # http://localhost:9770
```

## Features

- **Table** (`/`) — 14 sortable columns, multi-row select, full-text search, kind/status filters. Score column color-codes red→green.
- **Detail** (`/item/[id]`) — full metadata, scoring breakdown, monetization ideas, Google News feed, SWOT button.
- **SWOT** — one click runs 4 Claude Sonnet 4.6 agents in parallel (Strength / Weakness / Opportunity / Threat), each with its own quadrant-specific system prompt. Cached in Postgres.
- **Canvas** (`/canvas`) — selected items become React Flow nodes you can drag and connect; used for grouping marks into portfolio themes.
- **Score** — composite 0–100 from 6 weighted factors (distinctiveness 25, age-fit 20, recognition 15, domain 15, competing marks 15, monetization 10). Recompute via `POST /api/score`.

## Data sources

- USPTO TSDR (abandoned/cancelled trademark records)
- US Copyright Office public-domain records
- Google News RSS for per-item news (no API key needed)

Seed data is manually curated from these sources; scrapers can be added in `scripts/`.