▸ SEE IT IN ACTION
THE TERMINAL EXPERIENCE
Dark. Dense. Fast. Built for signal, not noise.
$ cat FEATURES.md
WHAT IT DOES
⬡
MULTI-AGENT INTELLIGENCE
6 philosophical AI agents analyze every asset simultaneously. Crack mapping, visibility analysis, narrative separation, and more.
◈
REAL-TIME MARKET DATA
Live price feeds, OHLCV data, and market metrics from Binance, CoinGecko, and on-chain sources.
⧖
AI CHAT AGENT
Built-in AI assistant with full context of your portfolio data. Ask anything about any asset, any time.
⊞
PER-AGENT MODEL CONFIG
Every agent runs its own model. Mix Claude, GPT, Gemini, and 280+ models across providers.
◎
CLASSIFICATION ENGINE
3-layer pipeline: Discovery, Classification (6 parallel agents), and Synthesis into structured verdicts.
▣
BLOOMBERG TERMINAL UX
Dark, dense, fast. Built for professionals who want signal, not noise. Every pixel earns its place.
$ cat ARCHITECTURE.md
HOW IT WORKS
LAYER 1
DISCOVERY
OpenCode agent in Docker — researches whitepapers, GitHub, socials, tokenomics. Returns structured context for all downstream agents.
LAYER 2 — 6 AGENTS IN PARALLEL
CRACK MAPPING
Structural fractures
VISIBILITY
Signal clarity
NARRATIVE SEP.
Signal vs noise
POWER VECTOR
Leverage mapping
PROBLEM RECOG.
Thesis validation
IDENTITY POL.
Community coherence
LAYER 3
SYNTHESIS
A 7th agent reads all 6 results and produces a unified verdict: category, conviction score, risk factors, and plain-language thesis.
→ ClassificationResult: category · conviction · risk_factors · thesis