Quickstart

Five minutes from zero to your first generated image. You need an account with a small prepaid balance and an API key.

1. Create an account and add balance

Sign up, then open Billing and add funds (minimum $5). Balance is debited per job at the published rates; failed and blocked jobs are refunded automatically.

2. Create an API key

In API keys, create a key. It is shown once — copy it into your secrets store and export it as FASTGEN_API_KEY. Keys look like fgc_live_….

3. Submit a job

Every generation endpoint returns 202 Accepted with a job. The optional Idempotency-Key header makes retries safe.

curl https://api.fastgencloud.com/v1/images/generations \
  -H "Authorization: Bearer $FASTGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: demo-0001" \
  -d '{
    "model": "fast-image",
    "prompt": "isometric illustration of a small coffee shop, warm light, clean vector style",
    "size": "1024x1024",
    "n": 1
  }'

4. Poll until it finishes

Fast image models usually finish in a few seconds when capacity is warm. Poll every 1–2 seconds, or skip polling entirely by passing a webhook_url (see Webhooks).

curl https://api.fastgencloud.com/v1/jobs/$JOB_ID \
  -H "Authorization: Bearer $FASTGEN_API_KEY"

5. Download the output

When status is succeeded, output.images[].url (or output.video.url) is a signed URL valid for up to 7 days and never past expires_at (your account’s retention window, 7 days by default). Store the file on your side if you need it longer.

{
  "id": "…",
  "object": "job",
  "status": "succeeded",
  "kind": "image.generation",
  "model": "fast-image",
  "input": { "prompt": "isometric illustration of …", "size": "1024x1024", "n": 1 },
  "output": {
    "images": [
      { "url": "https://…signed…", "width": 1024, "height": 1024, "mime": "image/png", "seed": 913 }
    ]
  },
  "usage": { "charged_micros": 4000, "charged_usd": 0.004, "gpu_ms": 1830 },
  "moderation": { "status": "passed" },
  "error": null,
  "created_at": "2026-08-17T10:00:00Z",
  "completed_at": "2026-08-17T10:00:04Z",
  "expires_at": "2026-08-24T10:00:04Z"
}

Image to video

Video works the same way: pass a start frame as a public https URL or a base64 data URI, an optional motion prompt, and a duration of 5 or 8 seconds. Billing is per second of output. Bring your own LoRAs by URL (up to 3): url is the high-noise file and low_url the optional low-noise pair — a singleurl is applied to both.

curl https://api.fastgencloud.com/v1/videos/generations \
  -H "Authorization: Bearer $FASTGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "image-to-video",
    "image": "https://example.com/still.jpg",
    "prompt": "slow camera push-in, gentle steam rising from the cup",
    "duration": 5,
    "loras": [
      { "url": "https://example.com/my-motion-lora-high.safetensors",
        "low_url": "https://example.com/my-motion-lora-low.safetensors",
        "strength": 0.9 }
    ]
  }'

Next steps

• GET /v1/models lists enabled models with supported sizes, durations and a price preview.
• Add webhook_url to any job and verify the signature — Webhooks.
• Handle 402, 422 and 429 — Errors and limits.
• Read what is and isn’t allowed — Content policy.
• Explore every field in the API reference.