Skip to main content

​
AI-Ready Documentation: Use Sorsa API Docs with LLMs

Sorsa’s documentation is built for use with AI assistants and large language models. You can feed the complete API reference directly into ChatGPT, Claude, Grok, Cursor, Copilot, or any other LLM to get accurate answers, generate integration code, and debug issues without manually reading through pages of docs.

​
The AI Docs File

We maintain a single, machine-readable file, sorsa_api_v3_documentation_en.txt, that contains the full Sorsa API specification: every endpoint, parameter, response schema, and authentication detail in one place. https://api.sorsa.io/v3/static/sorsa_api_v3_documentation_en.txt The whole file is compact (well under 20,000 tokens), so it fits comfortably in the context window of any modern LLM. Paste it into a conversation, attach it as a file, or add it to your project context, and the AI will have complete knowledge of the Sorsa API.

​
How to Use It

ChatGPT / Claude / Grok / Gemini: Copy the contents of the file (or attach it as a file) at the start of your conversation, then ask questions like “Write a Python script that searches for tweets about AI and exports them to CSV” or “How do I paginate through all followers of an account?” Cursor / Copilot / Claude Code / Codex / Windsurf: Add the file to your project context or reference it in your prompt. The AI will generate code that uses correct endpoint URLs, parameters, headers, and response parsing, without hallucinating field names or inventing endpoints that do not exist. Custom GPTs and assistants: Upload the file as a knowledge file when building a custom GPT or AI assistant that needs to work with X/Twitter data via Sorsa.

​
What’s Inside

The file is a structured reference for the entire API:
  • The base URL, authentication method, and common error codes
  • Shared data models (User, Tweet, Follower, Space, and more) with every field name and type
  • All endpoint paths grouped by category, each with its HTTP method and its required and optional parameters
  • Example requests and responses for every endpoint
  • Pagination behavior, sorting options, per-request limits, and date formats
  • Search syntax support, including common operators and a link to the complete operator reference
  • Edge cases and special response formats, such as the field-name differences between Tweet and Article objects (for example, view_count vs views_count)
It is kept up to date as the API evolves, so you always get the current specification.

​
Why This Matters

LLMs are only as good as the context they have. Without accurate API documentation, an AI assistant will guess at parameter names, invent response fields, or use outdated patterns. By providing the AI docs file, you get code that works on the first try instead of debugging AI-generated hallucinations.

​
Next Steps

To run the code your AI assistant generates, you need an API key. Every new account includes 100 free requests with no credit card required, so you can test generated code right away.