Yapper

A leaderboard dashboard ranking 101 indie tech founders by how much they've been yapping on X. Built with Next.js, Tailwind, and SQLite, styled after the Stanley design system.

How it works

• data/founders.json is the seed list (handle, name, product, tier, approximate followers) sourced from the curated spreadsheet.
• npm run refresh hits the X API v2 with a read-only bearer token: it batch-resolves all handles via GET /2/users/by (profile image, follower count, lifetime tweet count), then pulls each founder's last 30 days of posts via GET /2/users/:id/tweets and stores a dated snapshot in SQLite (data/yapper.db).
• The dashboard ranks founders by Yap Score: originals count 1.0, replies 0.5, retweets 0.25. Toggle between 7-day and 30-day windows.
• Snapshots accumulate over time, which powers the "biggest riser" stat (needs at least two refreshes).

Getting started

npm install

# Option A: demo data, no API token needed
npm run mock

# Option B: real data
cp .env.example .env.local   # then paste your X_BEARER_TOKEN
npm run refresh

npm run dev                  # http://localhost:3000

The bearer token comes from a read-only app in the X developer console. The refresh script handles rate limits by waiting and retrying, so a full run over 101 accounts may take a few minutes on the basic tier.

Scripts

Command	What it does
npm run dev	Start the dashboard
npm run refresh	Pull live data from the X API and snapshot it
npm run mock	Seed deterministic demo data (replaces existing snapshots)
node scripts/csv-to-seed.mjs <csv>	Regenerate data/founders.json from a spreadsheet export

Roadmap

• Threads will be added to this later.

GrayPass monitoring (time-boxed)

The site can run a two-week passive data-collection study for GrayPass: the drop-in SDK loads in monitor mode, asks each visitor for consent, and then streams anonymized interaction-pattern frames (typing rhythm, mouse movement, scrolling — never text or page contents) to the GrayPass API. Frames are observation-only (purpose="enroll", research_optin=true); GrayPass pseudonymizes them server-side into its research_donations training corpus. No overlays, no login gating — visitors who decline are never re-prompted.

Enable it by setting all four env vars (see .env.example): GRAYPASS_API_BASE, GRAYPASS_PUBLISHABLE_KEY, GRAYPASS_SECRET_KEY, and GRAYPASS_COLLECT_UNTIL (14 days after launch). The cutoff is enforced in three places — the layout stops injecting the script, /api/graypass/token refuses to mint tokens, and the SDK's data-until stops a running loop — so the study ends on time even without a redeploy.

On the GrayPass deployment itself, two settings must be in place or the data is silently discarded / blocked:

• GRAYPASS_RESEARCH_DONATIONS=1 — otherwise the research opt-in is ignored and nothing is persisted for training.
• GRAYPASS_ALLOWED_ORIGINS must include https://indiehackers.getstanley.ai — otherwise the browser's cross-origin frame posts fail CORS.

Structure

• app/ — Next.js App Router pages (server-rendered leaderboard)
• components/ — stat cards, top-3 cards, table, window toggle
• lib/ — SQLite access, leaderboard queries, X API client, Yap Score
• scripts/ — refresh, mock seeding, CSV conversion
• data/ — committed seed JSON; the SQLite database lives here too (gitignored)