Skip to main content
Market-intelligence reads are powered by , running as a self-hosted fork that generates fresh reads on a cron and commits them to a public repo. Protos fetches and caches them, so these answers are instant rather than waiting on a live model call.

DeFi read

Narrative map

A read on what story crypto Twitter/CT is currently telling itself — which sectors, which names, which direction sentiment is leaning.

Top DeFi protocols

Fear / divergence read

Flags when market sentiment and price action are telling two different stories.

Token pick

One scored pick a day — a 7-day dedup gate prevents the same coin from being re-served, and each candidate gets a 0–10 multi-signal score resulting in a HIGH, MEDIUM, or honest SKIP conviction call. A skip is a real, expected outcome on days nothing clears the bar — this isn’t a slot machine that always returns a pick.

Picks scorecard

A running scorecard of every past pick — win/hold/loss classification and hit rate, updated weekly. No cherry-picked dates.

Why cache-first matters

Because these reads come from a scheduled cron rather than a live model call at request time, they’re free to serve and consistently fast — on the web app and in headless/text mode (iMessage, Telegram) alike.