LLM Calls

Make LLM calls through the unified Zihin API.

Public Endpoint (API Key)

POST https://llm.zihin.ai/api/v3/llm/public/call

curl -X POST https://llm.zihin.ai/api/v3/llm/public/call \
  -H "X-Api-Key: zhn_live_xxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Explain quantum computing in simple terms",
    "model": "auto"
  }'

Authenticated Endpoint (JWT)

POST https://llm.zihin.ai/api/v3/llm/call

curl -X POST https://llm.zihin.ai/api/v3/llm/call \
  -H "Authorization: Bearer <jwt-token>" \
  -H "x-tenant-id: <uuid>" \
  -H "x-agent-id: <uuid>" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Explain quantum computing",
    "model": "auto"
  }'

Request Parameters

Parameter	Type	Required	Default	Description
query	string	Yes	—	The prompt
model	string	No	"auto"	Model ID or "auto"
provider	string	No	auto-detected	Provider name
messages	array	No	—	Conversation history
temperature	number	No	0.7	Creativity (0-2)
max_tokens	number	No	4096	Max response tokens
stream	boolean	No	false	Enable streaming

Conversation History

Send previous messages for multi-turn conversations:

{
  "query": "What about its applications?",
  "model": "auto",
  "messages": [
    {"role": "user", "content": "What is quantum computing?"},
    {"role": "assistant", "content": "Quantum computing uses quantum bits..."}
  ]
}

Response

{
  "success": true,
  "response": "Quantum computing uses quantum bits (qubits)...",
  "model": "gpt-4.1-mini",
  "provider": "openai",
  "usage": {
    "input_tokens": 15,
    "output_tokens": 120,
    "total_tokens": 135
  },
  "cost": 0.00045,
  "latency_ms": 850
}