Overview

Getting Started

Your business gets a custom AI model — trained on your use case, optionally grounded in your own documents — delivered through one simple REST API.

Prerequisites

An Intgr8AI account with API access (we set this up with you)
Your API key — see Step 1 below
Any HTTP client — curl, fetch, requests; no SDK required

Quickstart

Get your API key

Sign in at intgr8ai.com/dashboard
Open the API Usage section
Reveal and copy your key

Store your key as an environment variable and use it from your backend — never in client-side browser code.

export INTGR8AI_API_KEY="ik_live_your-key-here"

Verify your setup

Confirm the API is reachable with a quick health check (no auth needed).

curl https://intgr8ai-api-26046413214.us-central1.run.app/health

Make your first chat call

Send a conversation to POST /v1/chat. Your key already maps to your custom model — there's no model parameter to think about.

curl https://intgr8ai-api-26046413214.us-central1.run.app/v1/chat \
  -H "Authorization: Bearer $INTGR8AI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "What is your return policy?"}
    ]
  }'

Response:

200 OK

{
  "id": "chat_0936066790104ed6a089a0e9e99453e4",
  "message": {
    "role": "assistant",
    "content": "You can return any item within 90 days..."
  },
  "usage": { "input_tokens": 27, "output_tokens": 32 },
  "model": "support-bot"
}

Streaming

Add "stream": true to receive the reply token-by-token as server-sent events — delta chunks, a final usage event, then data: [DONE]. See the API Reference → Stream a chat completion for full examples.

Errors

Errors use a consistent envelope with a request_id (also on every response as the X-Request-Id header) you can reference with support.

Error envelope

{
  "error": {
    "type": "rate_limit_exceeded",
    "message": "Rate limit of 60 requests per minute exceeded.",
    "request_id": "49b8d78918eb49308429b4d08ec6c575"
  }
}

Status	Type	Description
401	invalid_api_key	Missing or invalid API key.
402	budget_exceeded	Credit balance is empty — top up to continue.
403	key_revoked	This key has been revoked.
404	not_found	Resource does not exist.
413	input_too_large	Message or document exceeds size limits.
422	invalid_request	Malformed request body or parameters.
429	rate_limit_exceeded	Slow down — check the Retry-After header.
502	upstream_error	Model provider failed; retried once. Safe to retry.

Rate limits & credits

Default limits are 60 requests/minute and 100,000 tokens/minute per key (custom limits available). On 429, wait the seconds given in Retry-After.

Billing is prepaid credits: requests burn credits based on tokens used. At zero balance the API returns 402 budget_exceeded until you top up. Track your balance with GET /v1/usage or in your dashboard.

Questions or need a key? Contact us — we set up your custom model and API access for you.