Power any Ayurveda app with AI Square

Models

Native AI² model family. Thinking depth controls how deeply the engine reasons before it responds — no third-party model names exposed in your product.

API keys

Generate keys to call AI² models from your backend. Early access is rolling out to partners.

Quick start

OpenAI-compatible chat completions. Scope classics with @Charaka in the user message or via metadata.scoped_works.

# List models
curl https://api.ai2.quaasx.com/v1/models

# Chat completion (non-stream)
curl https://api.ai2.quaasx.com/v1/chat/completions \
  -H "Authorization: Bearer $AI2_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ai2-ayu-pro",
    "messages": [
      {"role": "user", "content": "@Charaka @Sushruta Compare jwara nidana with citations"}
    ],
    "stream": false
  }'

# Streaming (SSE) — same endpoint with "stream": true
curl -N https://api.ai2.quaasx.com/v1/chat/completions \
  -H "Authorization: Bearer $AI2_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ai2-ayu-flash",
    "messages": [{"role": "user", "content": "What is jwara?"}],
    "stream": true
  }'

Models: ai2-ayu-flash, ai2-ayu-pro, ai2-ayu-max, ai2-ayu-god. Same-origin proxy path: /api/v1/* on the AI² site.