See the meta before it's the meta.
A completely autonomous pipeline that turns raw blockchain noise into readable narrative signals. Grouped by AI, not by hand.
How Tycho Reads The Chain
Instant Ingestion
Every launch is just three small signals — a name, a ticker, an image. Tycho reads all of them the moment they go live on Ponsfamily. No manual pasting required.
Vector Embedding
The raw text and image metadata is fed into Google Gemini, transforming the creative lore into multi-dimensional vectors.
Anomaly Detection
A DBSCAN clustering algorithm groups identical vectors together. When a group spikes abnormally fast, it flags it as a new narrative meta.
Active Narratives
Live clusters detected on the Ponsfamily ecosystem, mapped by spatial density.
One stream, pure signal.
The pipeline strips the noise and groups the remaining vectors into actionable metas.
Every token deployed.
Raw blockchain events contain thousands of duplicate or dead-on-arrival contracts.
Clustered narrative metas.
Only the tokens that share semantic vectors with high velocity are flagged.
Old way. New AI.
Finding metas used to mean staring at Telegram bots. Now it's fully autonomous.
Telegram Alpha
Scrolling through endless groups hoping to catch a narrative early.
DEX Screener
Setting volume filters and staring at new pairs for hours.
LLM Twitter Bots
Following specific influencers to trade based on their tweets.
Autonomous Radar
Vectors and clustering find the metas mathematically.
The pipeline compounds.
Every 5 minutes, the cron job executes the complete pipeline end-to-end.