For developers

Build trading agents on GT Protocol

Connect an AI agent to a real trading engine. GT Protocol ships an open Model Context Protocol server, so your model can build, backtest and run strategies from code or straight from an LLM client.

READCheck balances, list bots, read deals and positions
TESTRun backtests on real market history
TRADECreate, start, update and stop strategies
MANAGEOpen and close deals across CEX and DEX

What you get

MCP server

The GT Protocol API as Model Context Protocol tools. Point Claude (or any MCP client) at it and the model can check balances, run backtests and manage bots directly.

Trading API

The same engine the GT App runs on. Create strategies, launch deals and read results back across Binance and Hyperliquid from your own code.

Your agent, our execution

Skip wiring exchanges, signatures and order routing from scratch. Build the agent; GT Protocol handles execution on a layer that already trades real money.

From idea to live strategy

The same primitives the GT App uses are open to you: connect a venue, define a strategy, test it on real history, then launch, all programmatically.

01

Connect

Link a centralized exchange by API key or a Hyperliquid wallet, with trading-only access.

02

Backtest

Validate a strategy on real historical data before any capital is at risk.

03

Launch

Run it live and read deals, positions and results back through the API.

Get access

Building an agent or integration on GT Protocol? The MCP server is open source. Read the docs to wire it up, or message the team for API access.