One API, one key: the Watt AI analyst, REST market data (NEM + WEM), alerts and webhooks, and the grid carbon-intensity feed — all included. A dual-model architecture routes tool calls and synthesises analyst-grade answers. Ship it inside your product with one API call. Tiers are priced by call volume, never by feature.
Send a natural-language message. Watt decides which tools to call (prices, generation, carbon, constraints, PPA evaluation, demand response, fragility scores, and more), fetches live data, and synthesises a final analytical response.
curl -X POST https://gridiq.com.au/api/watt/query \
-H "Authorization: Bearer giq_..." \
-H "Content-Type: application/json" \
-d '{"message": "What is driving VIC prices right now?"}'Every response includes the tools consulted and your current usage:
{
"response": "VIC is trading around $112/MWh this afternoon. Two factors...",
"tools_used": ["get_current_prices", "get_generation_mix", "get_weather_context"],
"api_usage": {
"api_tier": "developer",
"month": "2026-04",
"used": 2341,
"cap": 25000,
"remaining": 22659
}
}Authenticated via API key (Bearer token). Rate-limited per key with atomic check-and-increment metering. Fails closed on database errors so you never get charged for calls that weren't honoured. IP rate limits stack on top as burst protection.
Model versions are pinned and the tool list is versioned via the response payload. Contract stability is prioritised: breaking changes require a major version bump and a three-month sunset of the prior version.