HTTP API Guide¶

Use LLM Council as an HTTP server with REST API and SSE streaming.

Installation¶

pip install "llm-council-core[http]"

Start Server¶

llm-council serve --port 8001

Endpoints¶

POST /v1/council/query¶

Submit a query to the council.

curl -X POST http://localhost:8001/v1/council/query \
  -H "Content-Type: application/json" \
  -d '{
    "query": "What are best practices for error handling?",
    "confidence": "balanced"
  }'

GET /v1/council/stream¶

Stream council events via SSE.

curl -N "http://localhost:8001/v1/council/stream?prompt=What+is+AI"

GET /v1/health¶

Health check endpoint.

curl http://localhost:8001/v1/health

SSE Events¶

Event	Description
`council.deliberation_start`	Council begins
`council.stage1.complete`	Initial responses ready
`model.vote_cast`	Model submitted ranking
`council.stage2.complete`	Rankings complete
`council.complete`	Final answer ready

Client Example (JavaScript)¶

const source = new EventSource('/v1/council/stream?prompt=...');

source.addEventListener('council.complete', (e) => {
  const result = JSON.parse(e.data);
  console.log('Answer:', result.stage3_response);
  source.close();
});