FAQ: AI

​

How do I get started with AI integration?

OneBalance offers multiple ways to integrate with AI tools. The easiest approach depends on your use case:

• For general LLM integration: Use our LLM feed files (llms.txt and llms-full.txt)
• For code editors like Cursor: Add our documentation directly to your editor’s indexing
• For advanced API access: Set up our MCP server for live OneBalance API functions

Start with the Building with LLMs guide to explore all options.

​

What’s the difference between llms.txt and llms-full.txt?

We provide two different feed files for different use cases:

• llms.txt - A concise, high-signal list of top-level documentation pages. Perfect for smaller models or when you need quick context building.
• llms-full.txt - Complete documentation content including all pages. Ideal for full-context indexing and AI training.

Both files are automatically updated and available at:

• https://docs.onebalance.io/llms.txt
• https://docs.onebalance.io/llms-full.txt

Learn more about LLM feed files.

​

How do I use OneBalance docs in my code editor?

For Cursor users:

1. Navigate to Cursor Settings > Indexing & Docs
2. Select “Add Doc” and paste: https://docs.onebalance.io/llms-full.txt
3. Use @docs -> OneBalance to reference docs in your code

You can also access any documentation page as plain markdown by adding .md to the URL (e.g., /getting-started/introduction.md).

See the complete code editor integration guide.

​

What is the MCP server and how do I set it up?

The Model Context Protocol (MCP) server connects OneBalance’s documentation and API directly to your AI tools. It provides:

• Documentation search across all OneBalance docs
• Live API access to OneBalance functions (balances, quotes, transactions, etc.)
• Direct integration with Cursor, Claude Desktop, and other MCP clients

Quick setup:

npx mint-mcp add onebalance-d7e5d4d0

The MCP server gives you 11 functions including documentation search and full OneBalance API access. Follow our complete MCP server setup guide for detailed instructions.

​

Can I test the OneBalance API without my own key?

Yes! We provide a public API key for testing purposes with limited usage:

42bb629272001ee1163ca0dbbbc07bcbb0ef57a57baf16c4b1d4672db4562c11

You can use this key when setting up the MCP server or making direct API calls. This lets you explore OneBalance functionality before getting your own API key.

For production use, you’ll need your own API key from your dashboard or by contacting support@onebalance.io.