Alphametrics
Crypto Twitter is full of "calls" — but who actually delivers, and who just uses their followers as exit liquidity? Alphametrics answers that on-chain: every call verified against the blockchain, with a formal definition of what counts as a win.
Next.js
Supabase
Solana
On-chain alpha. Verified.
Anyone can screenshot a winning call and delete the losers. Alphametrics removes the ambiguity by defining a win formally and checking it against the chain: a call counts only if the token gained at least 10% from the tweet, with the peak arriving at least three minutes after — so it was actually followable, not already gone.
Analyze any handle and it pulls their last ~3,200 tweets, extracts every Solana contract mentioned, and builds a verified performance history with time-to-peak and an equal-weight portfolio simulation.
A real-time signal engine on top of the chain.
Beyond historical verification, it watches the chain live:
- Helius webhooks and a standalone WebSocket worker classify wallet activity into signal types — whale buys, sniper alerts, dormant-wallet buys, bundle-exit detection and more, each with SOL thresholds.
- Analysis progress and live signals both stream to the frontend over Server-Sent Events.
- Multi-timeframe leaderboards (24h / 7d / 30d / 90d / all) run off tuned Postgres procedures across 15 migrations.
- The same SSE client is reused by Discord and Telegram bots — one code path, three surfaces — so /analyze streams live progress into a chat embed.
Planning something similar?
I design and ship AI agent systems, data platforms and full-stack products — from first idea to production.
Get in touch