Can I vibecode IdeaFast?

Build me a Reddit pain finder: a local CLI plus a small dashboard that reads public
Reddit and turns complaints into ranked pain themes with clickable evidence.

Functional & CLI Requirements:
- TypeScript on Node 22, SQLite via better-sqlite3, Hono for the dashboard. One repo, one `npm run scan` entrypoint. No accounts, no cloud, no telemetry.
- Input: subreddits.txt plus a timeframe flag (default 90 days). Fetch posts and top level comments from Reddit's public JSON endpoints, for example `https://www.reddit.com/r/<sub>/top.json?t=year`.
- Rate Limiting & Caching: Enforce strictly one request every 2 seconds. Use a real, descriptive User-Agent. Cache every raw JSON response in SQLite so re-runs cost nothing. No OAuth, no logged-in scraping.
- Pre-filtering: Filter down to complaint-shaped text using highly optimized, cheap regexes (e.g., "I hate", "why is there no", "wasted hours", "workaround", "gave up on", "frustrating") before spending a single API token.

AI Pipeline & Clustering:
- Classification: Pass surviving texts to an LLM (e.g., Claude with ANTHROPIC_API_KEY in .env, batched, cached by content hash) to determine: `is_pain` (boolean), `severity` (1 to 5), and a `one_line_summary`.
- Embeddings & Grouping: Embed the generated summaries using a lightweight local embedding model or API, group them at a cosine similarity threshold above 0.82.
- Naming & Evidence: Prompt the LLM to name each cluster and extract the 5 strongest verbatim quotes with permalinks. Never paraphrase a quote; evidence must be exactly as written and clickable.
- Scoring: Score each cluster using the formula: `frequency * mean_severity * recency_decay`. Persist these scores in the database so subsequent scans can track movement over time.

Dashboard & Visualization:
- Serve a fast, client-side rendered dashboard via Hono on `localhost:3000`.
- UI features: List ranked clusters, allow expanding to view verbatim quotes with permalinks, filter by specific subreddit, and provide a one-click CSV export of the data.
- Ensure the interface is highly readable, using color gradients to represent severity and trend arrows for score changes.

DevOps & Documentation:
- README: Document how to choose subreddits, the critical importance of the rate limit rule (and why breaking it gets you blocked by Reddit), and provide a rough token cost estimate per 1000 comments.
- Out of scope: Sources other than Reddit, idea generation, and cross-scan dedupe. Focus exclusively on getting one subreddit list to produce reliable, trusted pain clusters first.