Skip to main content
Dakota has made its documentation AI-friendly through multiple mechanisms designed to help Large Language Models understand and integrate with the platform.

​
LLM Feed Files

To help LLMs stay current on how Dakota works, we expose two continuously updated files for ingestion:
  • llms.txt - A concise, high-signal list of top-level docs pages, great for smaller models or quick context building.
  • llms-full.txt - A more exhaustive listing that includes nearly all pages, ideal for full-context indexing.
You can regularly ingest these URLs into your custom GPTs or other LLM apps to ensure Dakota-specific questions are grounded in accurate technical detail. The documentation supports “contextual” features allowing you to:

​
Export as Markdown

Export any Dakota documentation page as Markdown for:
  • Custom GPT training data
  • Internal knowledge bases
  • Team documentation
  • Offline reference

​
AI Chat Integration

Launch pre-loaded chat sessions with Claude or ChatGPT for specific documentation pages. This enables:
  • Instant troubleshooting
  • Code generation with proper context
  • Deeper topic exploration
  • Interactive learning

​
Use Cases

Troubleshooting: Open a docs page about webhooks, click “Ask Claude”, and get immediate help with your specific webhook implementation issue. Code Generation: Load the API reference page, start a chat, and generate production-ready code that follows Dakota’s best practices. Learning: Explore complex topics like transaction flows by chatting with an AI that has full context of Dakota’s documentation.

​
Best Practices

​
Regular Ingestion

For custom GPTs or internal tools:
  • Fetch llms.txt or llms-full.txt regularly (daily or weekly)
  • Update your knowledge base with the latest documentation
  • Ensure accurate, current technical information

​
Context Management

  • Use llms.txt for general queries and overviews
  • Use llms-full.txt when detailed implementation guidance is needed
  • Combine with live API testing for verification

​
Security Considerations

  • Never share API keys with AI assistants
  • Use sandbox credentials when generating code examples
  • Review AI-generated code before production deployment
  • Verify security recommendations against official docs

​
Getting Started

  1. Choose Your Integration Method
    • Quick start: Use llms.txt with your AI assistant
    • Full context: Ingest llms-full.txt into custom GPTs
  2. Test Your Setup
    • Ask basic questions about Dakota concepts
    • Request code examples for common operations
    • Verify responses against official documentation
  3. Build with Confidence
    • Generate boilerplate integration code
    • Get instant answers to API questions
    • Troubleshoot issues with AI assistance

​
Resources

​
Support

Need help integrating Dakota documentation with your AI tools? Contact our support team at dakota.io/talk-to-sales for assistance.