API

import { QuiverAI } from "@quiverai/sdk";

const client = new QuiverAI({
  bearerAuth: process.env["QUIVERAI_API_KEY"],
});

const logo = await client.createSVGs.generateSVG({
  model: "arrow-1.1",
  prompt: "A logo for the next AI design startup",
  instructions: "Use clean geometry, balanced spacing, and production-ready SVG structure.",
});

process.stdout.write(`${JSON.stringify(logo, null, 2)}\n`);

Use HTTP directly from any language.

curl --request POST \
  --url https://api.quiver.ai/v1/svgs/generations \
  --header 'Authorization: Bearer <QUIVERAI_API_KEY>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "arrow-1.1",
  "prompt": "A logo for the next AI design startup",
  "instructions": "Use clean geometry, balanced spacing, and production-ready SVG structure.",
  "n": 1,
  "stream": false
}
'

A successful non-streaming response returns request metadata, one or more SVG payloads, and the request credit debit:

{
  "id": "resp_01J9AZ3XJ7D5S9ZV2Q5Z8E1A4N",
  "created": 1704067200,
  "data": [
    {
      "svg": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M12 2l8 20H4z\"/></svg>",
      "mime_type": "image/svg+xml"
    }
  ],
  "credits": 20
}

The deprecated usage token fields may still appear for compatibility. These SVG routes use fixed-credit models, so use credits and pricing_credits for billing behavior. For token-usage models on /v1/responses, read the model’s billing.rates instead.

Failures return a JSON error payload:

{
  "status": 429,
  "code": "rate_limit_exceeded",
  "message": "Rate limit exceeded",
  "request_id": "550e8400-e29b-41d4-a716-446655440000"
}

• 401 Unauthorized: API key missing or invalid, or the organization could not be resolved for billing. Machine-readable codes include invalid_api_key and unauthorized.
• 402 Payment Required: insufficient credits.
• 403 Forbidden: frozen accounts use the account_frozen code.
• 429 Too Many Requests: back off and, when Retry-After is present, retry after the indicated number of seconds. X-RateLimit-Limit and X-RateLimit-Remaining are counts in the current binding window. X-RateLimit-Reset is that window’s Unix reset timestamp in milliseconds when available; rolling token-capacity rejections omit it. Positive token-capacity ceilings report their retry interval through Retry-After, while zero ceilings omit retry timing. X-RateLimit-Scope, optional X-RateLimit-Subject, and X-RateLimit-Dimension (request_rate, operation_throughput, input_token_rate, or output_token_rate) name the binding ceiling.

Billing model:

• These SVG operations debit pricing credits from your balance; amounts are per fixed-credit model (see that model’s pricing_credits from GET /v1/models).
• Generations debit n × svg_generate credits on success (n defaults to 1).
• Vectorizations debit svg_vectorize credits per successful request.
• Token-usage models on /v1/responses settle measured tokens using the model’s billing.rates; they do not expose pricing_credits.