Using AI with 0x: an overview

0x provides several ways to work with AI tools and agents — both to help developers build integrations faster, and to let AI agents trade autonomously. This article summarizes what's available. For setup commands, configuration, and code, each section links to the canonical docs.

The starting point

The Using AI with 0x page is the hub. 0x Docs includes built-in AI integrations, 0x skills, and an official MCP server so you can work with up-to-date documentation directly inside your AI tools and editor. Every documentation page also has a page-actions menu with options to copy the page in an AI-optimized format, view its raw Markdown, ask an in-docs AI a question, or open the page directly in ChatGPT or Claude.

Agent Skills

Agent Skills give an AI coding agent built-in knowledge of the 0x APIs. Once installed, your agent knows every endpoint, authentication method, and integration pattern for token swaps, and it's compatible with Claude Code, Cursor, GitHub Copilot, and any agent that supports the Agent Skills specification. The skill covers swap flows, gasless transactions, authentication, error handling, and chain-specific details. This is the recommended starting point if you're actively building a swap integration in your editor. See Agent Skills.

MCP Server

The 0x MCP (Model Context Protocol) Server connects your editor or AI assistant to 0x documentation. Instead of copy-pasting documentation into prompts, your AI assistant can retrieve the right guides, API references, and examples automatically, which keeps it current even when docs have changed since its training cutoff. You can then ask it to explain concepts, generate implementation code for a given chain and language, or help debug integration errors. See MCP Server Setup.

Skills vs. MCP

They solve different problems and work well together: Agent Skills embed API knowledge so the agent can write correct code, while the MCP server lets the agent fetch live docs and references on demand. Use both for the best experience.

Autonomous Agent Payments

This is aimed at AI agents that need to call APIs on their own, without human-managed keys or billing. Two open standards — x402 and MPP (Machine Payments Protocol) — embed micropayments directly into HTTP requests, and 0x swap endpoints are available through both via the Alchemy AgentPay proxy, where agents pay $0.01 per request in USDC with no 0x API key required — payment is the authentication. x402 settles in USDC on Base or Solana, while MPP settles in USDC.e on Tempo Mainnet. See Autonomous Agent Payments.

AI Project Showcase

If you want to see what others have built, the AI Project Showcase features developer frameworks and consumer apps using 0x — including agent frameworks like ElizaOS and Goat-SDK, and consumer products like BankrBot, Cortex, and NANI. You can also request to have your own project featured.

Where to go next

This is a conceptual overview. For installation commands, MCP configuration, payment-protocol endpoints, and code samples, follow the links above into the 0x documentation, which is always the source of truth.