@wzrd_sol/signal
v0.1.3
Published
Which AI model should I use? Real-time velocity signals for 100+ models across 4 platforms.
Maintainers
Readme
@wzrd_sol/signal
Which AI model should I use? Real-time velocity signals for 100+ models.
npm install @wzrd_sol/signalimport { pick, shortlist, prices } from "@wzrd_sol/signal";
// Best model for code tasks right now
const model = await pick("code");
// → "Qwen/Qwen3.5-9B"
// Top 5 models for chat
const top5 = await shortlist("chat", 5);
// → [{ model, score, trend, platform, ... }, ...]
// CCM and vLOFI prices
const p = await prices();
// → { ccm: { price_usd: 0.000789 }, vlofi: { price_usd: 0.001764 } }What it does
WZRD tracks adoption velocity of open-source AI models across HuggingFace, GitHub, OpenRouter, and ArtificialAnalysis. This package gives you that signal as a simple function call.
pick(task)— best model for a task ("code", "chat", "reasoning")shortlist(task, n)— top N models ranked by velocitypickDetails(task)— full signal with trend, confidence, platformfetchSignals()— raw feed dataprices()— CCM/vLOFI token prices
Zero dependencies
Works in Node.js, Deno, Bun, and browsers. Uses native fetch.
On-chain oracle
9 Switchboard feeds on Solana mainnet (7 velocity + 2 price) publish the live oracle set on-chain. Any Solana program can read model momentum without an API call.
