Build your first Wan 3 generation

Base URL

Send every public API request to this versioned base URL:

https://api-platform.gensta.ai/v1

The /v1 segment versions the public HTTP contract. Compatible additions can ship inside v1; a breaking contract requires a new major version.

Before you start

1. Get a project API key.Keep it on your server. Never expose it in browser or mobile code.
2. Read the live catalog.GET /v1/models needs no key and is the source of truth for availability, capabilities, and customer rates.
3. Create one logical generation.Use a stable Idempotency-Key so a lost response can be retried safely.
4. Poll by ID.Video is asynchronous; stop when the generation reaches a terminal status.

Create a generation

export GENSTA_API_KEY="your-secret-key"

curl --silent --show-error --fail-with-body \
  -X POST "https://api-platform.gensta.ai/v1/generations" \
  -H "Authorization: Bearer $GENSTA_API_KEY" \
  -H "Idempotency-Key: order-8493-shot-1" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "wan-3.0",
    "mode": "text_to_video",
    "variant": "standard",
    "parameters": {
      "resolution": "720p",
      "duration_seconds": 5,
      "ratio": "16:9",
      "audio": true,
      "prompt_extend": true
    },
    "input": {
      "prompt": "A quiet lake at sunrise, cinematic aerial shot"
    }
  }'

A valid create returns HTTP 202 with an id, status: queued, and a USD price hold snapshot. Save the ID before doing anything else.

Poll for a result

curl --silent --show-error --fail-with-body \
  "https://api-platform.gensta.ai/v1/generations/$GENERATION_ID" \
  -H "Authorization: Bearer $GENSTA_API_KEY"

Start after about five seconds, then back off to 10–15 seconds with jitter. On succeeded, download artifacts[].download_url promptly. The URL is short-lived; a later GET returns a newly signed URL for the same artifact.