What is Pyth?

Pyth is an oracle. That means it gives smart contracts something they cannot fetch themselves: a fresh price. The thing that makes Pyth different from the rest is who is doing the publishing. Instead of a middle-man relaying data from somewhere else, the trading firms and exchanges that see the prices in their own books push them straight into the network.

Today that pipe carries prices for over 3,059 markets — crypto, stocks, FX, metals, commodities, ETFs — and pours them onto 100+ blockchains. Around 1,000+ partners read from it. Perp DEXes, lending markets, RWA platforms, prediction games, on-chain insurance. If a smart contract anywhere on a major chain needs a price, there is a good chance it gets it from Pyth.

The 60-second pitch

• First-party data. 138+ publishers including Jane Street, Cboe, Binance, OKX, Wintermute, Two Sigma, Optiver. They sign the price with their own keys. No node operator stands between them and you.
• Pull oracle. Prices live off-chain in a service called Hermes. When your contract needs one, it pulls a signed update and posts it on-chain. You pay only for the updates you use.
• Sub-second freshness. Updates aggregate every 400 milliseconds on Pyth Core. Pyth Pro pushes that to ~1ms for trading-grade venues.
• Confidence intervals. Every Pyth price comes with a conf value — basically an error bar. Lending markets use it to be cautious; traders use it to detect when feeds are noisy.
• Permissionless on Core. Anyone can read any Core feed. You do not need to apply or sign a contract. Pro is a paid subscription.

Where it came from

Pyth shipped on Solana in 2021 because that chain had the throughput to handle 400ms updates without dying. The team kept the publishers on a dedicated chain — Pythnet — and used Wormhole to relay signed prices outward. That move is what turned Pyth from a Solana-only project into the cross-chain oracle it is today.

The bet paid off. As Ethereum L2s and the new wave of app-chains went live, every one of them needed prices. Pyth was the only oracle ready to drop into a contract on day one.

Why first-party data matters

Most oracles are node operator oracles. A network of nodes scrapes prices from public APIs — usually third-party aggregators that themselves scrape exchanges — then post a median on-chain. Two problems with that:

1. You are reading a copy of a copy. By the time the price is on-chain, it has been through three layers of polling. For lending that is fine. For perpetuals it is not.
2. The data source is paid in goodwill. The exchanges that produced the prices in the first place get nothing. Pyth flipped that — publishers are paid in PYTH to publish, and they have a direct stake in keeping their feeds accurate.

What this guide will cover

The next ten chapters work outward from the basics. Chapter 2 explains how a price actually moves from a trading firm to your contract. Chapter 3 dissects what a price looks like on-chain. After that we walk each product Pyth ships — Pyth Core, Pyth Pro, Entropy, Express Relay, Marketplace — then the network around it, the use cases, the token, and a glossary you can keep open in a second tab.

Pick the path that fits. If you build, jump to chapter 3. If you research, read it straight through. If you only need a word defined, the glossary is one click away.