AI Agents·2026

Confucius

Confucius is a research agent for the Solana trenches, dressed as a retro CRT terminal. Ask it about a token, a wallet or an influencer in plain language, and it fans out across ~20 tools in parallel — then renders the answer as living UI, not a wall of text.

Claude

Solana

Next.js

TypeScriptVisit site

~20

tools

tool families

reasoning steps max

∥

parallel tool calls

Confucius terminal: an amber CRT-style interface listing token analysis, wallet tracking, discovery and social intel capabilities — The CRT terminal — ask naturally, get structured intelligence back.

01 — The Idea

Research, no trading risk.

After building a run of trading agents that touch real wallets, Confucius is the pure-research cut: all the on-chain and social intelligence, none of the execution risk. The persona is Confucius — ancient wisdom meets live market data — and everything happens through natural language in the terminal.

02 — Under the Hood

Parallel tools, generative UI.

Two things make it feel fast and legible:

Around 20 tools span five families — token analysis (risk, holders, metadata, charts), wallet analysis (PnL, history), discovery (trending, graduating, new), social intel (influencer track records), and trading info (prices, quotes).
The system prompt explicitly instructs parallel tool calls — when several independent data points are needed, they're fetched at once, not in sequence.
Each tool renders its own React component — a risk meter, a holder table, an influencer track-record card — so answers come back as interactive UI, not markdown, all inside the CRT aesthetic.

Planning something similar?

I design and ship AI agent systems, data platforms and full-stack products — from first idea to production.

Get in touch