vidu/q4-preview/reference-to-videoVidu Q4 Preview reference-to-video builds one scene from 1 to 15 reference images and up to 3 voice clips, in 16:9, 9:16, 1:1, 3:4 or 4:3. Clips run 3 to 16 seconds at 540p to 4k with generated sound. It is lightly filtered: from the Moderate level up it tones down or refuses most NSFW requests.
Filtering differs by endpoint: Image to video — Uncensored, Video extension — Uncensored, Reference to video — Light.
Examples use the selected endpoint’s public parameter definitions. Replace the API key and example media, then send the request from your server.
SDKs & integrationshttps://api.spicyapi.ai/api/v1/jobs/createTaskKeep the API key in the server environment variable SPICY_API_KEY. Use Bearer authentication and Content-Type: application/json for JSON requests.
Authorization: Bearer $SPICY_API_KEY# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "vidu/q4-preview/reference-to-video",
"input": {
"aspect_ratio": "16:9",
"duration_seconds": 5,
"prompt": "A slow aerial shot drifting over a misty pine forest at sunrise",
"reference_image_urls": [
"https://example.com/reference.png"
],
"resolution": "720p"
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"Generated from this model’s input schema — the same one the API validates every request against.
Place these fields inside the request body’s input object. Names and types match the API. Do not assume a default when none is specified.
reference_image_urlsarray<string>RequiredRequired, 1 to 15 images of people, outfits, products or places. A request without any is rejected. The order matters because the prompt refers to them by position. Does not change the price.
promptstringRequiredRequired, up to 20,000 characters. Describe the scene and refer to references by position, such as "the woman from the first image". Put spoken lines in double quotes. Does not change the price.
seedintegerOptionalKeeps runs comparable while you change one thing at a time. It does not promise a frame-identical clip. Does not change the price.
generate_audiobooleanOptionalOn by default: dialogue, effects and ambience are generated with the picture. Turn it off for a clip with no audio track. Does not change the price.
duration_secondsintegerOptionalAffects priceWhole seconds, 3 to 16, default 5. Billed per second together with the resolution.
resolutionstringOptionalAffects price540p, 720p (the default), 1080p, 2k or 4k. This sets the price per second: 2k costs more than 1080p, and 4k about twice 2k. 2k and 4k return 10-bit HEVC files that some browsers cannot play inline.
"540p""720p""1080p""2k""4k"aspect_ratiostringOptional16:9 (the default), 9:16, 1:1, 3:4 or 4:3. Does not change the price.
"16:9""9:16""1:1""3:4""4:3"reference_audio_urlsarray<string>OptionalOptional, up to 3 voice clips (Vidu asks for MP3, 3 to 12 seconds each). A clip shapes the voice but not the words: write the line in the prompt and say it is in the voice from the reference audio. Does not change the price.
{
"additionalProperties": false,
"properties": {
"aspect_ratio": {
"default": "16:9",
"description": "Output aspect ratio.",
"enum": [
"16:9",
"9:16",
"1:1",
"3:4",
"4:3"
],
"type": "string",
"x-ui": {
"order": 80,
"widget": "select"
}
},
"duration_seconds": {
"default": 5,
"description": "Output length in whole seconds, from 3 to 16. Billed per second.",
"maximum": 16,
"minimum": 3,
"type": "integer",
"x-ui": {
"advanced": false,
"affects_price": true,
"order": 70,
"unit": "s",
"widget": "slider"
}
},
"generate_audio": {
"default": true,
"description": "Generate sound with the picture: dialogue, effects and ambience. Off gives a silent clip at the same price.",
"type": "boolean",
"x-ui": {
"affects_price": false,
"order": 60,
"widget": "switch"
}
},
"prompt": {
"description": "Describe the scene and refer to each reference by its position: \"the woman from the first image\", \"the robe from the second image\".",
"maxLength": 20000,
"minLength": 1,
"type": "string",
"x-ui": {
"order": 100,
"primary": true,
"reference_tokens": {
"reference_audio_urls": "Audio {index}",
"reference_image_urls": "Image {index}"
},
"widget": "textarea"
}
},
"reference_audio_urls": {
"description": "Up to three reference audio clips (MP3, 3 to 12 seconds each) for a voice or a sound the clip should follow.",
"items": {
"contentMediaType": [
"audio/mpeg",
"audio/wav"
],
"format": "uri",
"type": "string"
},
"maxItems": 3,
"minItems": 1,
"type": "array",
"x-ui": {
"affects_price": false,
"order": 86,
"widget": "multi-upload"
}
},
"reference_image_urls": {
"description": "One to 15 reference images: people, outfits, products or places, in the order your prompt refers to them.",
"items": {
"contentMediaType": [
"image/jpeg",
"image/png",
"image/webp",
"image/gif"
],
"format": "uri",
"type": "string"
},
"maxItems": 15,
"minItems": 1,
"type": "array",
"x-ui": {
"affects_price": false,
"order": 90,
"widget": "multi-upload"
}
},
"resolution": {
"default": "720p",
"description": "Output resolution. 2k and 4k come back as 10-bit HEVC.",
"enum": [
"540p",
"720p",
"1080p",
"2k",
"4k"
],
"type": "string",
"x-ui": {
"advanced": false,
"affects_price": true,
"order": 75,
"widget": "radio"
}
},
"seed": {
"description": "Random seed for controlled comparisons, not a frame-identical guarantee.",
"maximum": 2147483647,
"minimum": 0,
"type": "integer",
"x-ui": {
"order": 10,
"widget": "number"
}
}
},
"required": [
"prompt",
"reference_image_urls"
],
"type": "object",
"x-pricing": {
"billingMode": "variant_matrix",
"variantFields": [
"resolution"
]
}
}HTTP 202 means the task was accepted, not completed. Save the returned taskId and query that task. Do not create another task while waiting for its result.
HTTP/1.1 202 Accepted
Location: /api/v1/jobs/recordInfo?taskId=job_01k3m8x9q2z4v7n5p6r8s0t1w3
{
"code": 200,
"msg": "success",
"request_id": "req_example_create",
"data": {
"taskId": "job_01k3m8x9q2z4v7n5p6r8s0t1w3",
"state": "queued",
"estimatedCost": "0.750000"
}
}The accepted response includes the estimate so your app can show the same amount in task history and billing.
curl --get "https://api.spicyapi.ai/api/v1/jobs/recordInfo" \
--data-urlencode "taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"Start polling at about 2 seconds and gradually increase the interval. Stop at a terminal state. After a connection failure, query the original taskId. Retry creation only with the original idempotency key and body.
import { verifyWebhook } from "@spicyapi/sdk";
// express.raw keeps the bytes we signed; express.json never matches.
app.post("/hooks/spicy", express.raw({ type: "application/json" }), (req, res) => {
const { payload } = verifyWebhook({
rawBody: req.body,
timestamp: req.header("X-Webhook-Timestamp"),
signature: req.header("X-Webhook-Signature"),
payloadVersion: Number(req.header("X-Webhook-Payload-Version")),
secret: process.env.SPICY_WEBHOOK_SECRET,
});
res.status(200).end();
void handleFinishedTask(payload);
});Pass callBackUrl when you create the task and the finished record is POSTed to you — no polling loop at all. Every delivery carries X-Webhook-Timestamp, X-Webhook-Signature and X-Webhook-Payload-Version. The signature is a base64 HMAC-SHA256 over taskId.timestamp.sha256(body), so it must be checked against the raw bytes: parsing the JSON and re-serialising it changes them and the check will fail. Failed deliveries are retried, so the same taskId can arrive twice — make the handler idempotent. Answer 2xx first and do the work afterwards; a slow handler is treated as a failed delivery.
Consume output only when state is succeeded. On failure, use errorCode, errorMessage and request_id to investigate. Example costs and states do not represent an actual bill.
{
"code": 200,
"msg": "success",
"request_id": "req_example_record",
"data": {
"taskId": "job_01k3m8x9q2z4v7n5p6r8s0t1w3",
"model": "vidu/q4-preview/reference-to-video",
"state": "succeeded",
"output": {
"assets": [
{
"key": "tasks/2026/08/28/job_01k3m8x9q2z4v7n5p6r8s0t1w3/3f7a1c9e5b2d48f0a6c1e3d5b7f9a1c3.mp4",
"mime": "video/mp4",
"url": "https://example.r2.cloudflarestorage.com/results/job_01k3m8x9q2z4v7n5p6r8s0t1w3/result.mp4?X-Amz-Signature=SIGNATURE_FROM_RESPONSE",
"expiresAt": "2026-08-28T12:20:08.000Z"
}
]
},
"cost": "0.750000",
"settled": true,
"createdAt": "2026-08-28T12:00:00Z",
"completedAt": "2026-08-28T12:00:30Z"
}
}For each available asset, open or download assets[].url directly without an API Authorization header. The signed URL expires at expiresAt; poll the task again to refresh it while the asset remains within its retention period. A pending or unavailable asset has no URL. Treat signed URLs as private.
Use SpicyAPI in your app, terminal, AI assistant or visual workflow. Choose a setup guide below.
You do not have to write any of the code above. Connect the MCP server once and any agent that speaks MCP — Claude Code, Codex, Cursor, Windsurf, VS Code, Gemini CLI, Cline, LibreChat — can run this model by name, or run it from your shell with a single command.
Use vidu/q4-preview/reference-to-video to make a five-second shot of a neon-lit rooftop at dusk, slow push-in, and save it to ./out/.
The agent looks the model up in the live catalogue, uploads the local file for you, then shows the exact quote and its maximum charge and waits for your confirmation. Nothing is billed until you say yes.
claude mcp add spicyapi -e SPICY_API_KEY=$SPICY_API_KEY \
-- npx --yes --package=@spicyapi/mcp spicyapi-mcp
npx skills add spicyapi-ai/spicy-skillThe first line gives the agent the tools; the second gives it the Skill — the operating rules for them, so it quotes before it bills, uploads local files instead of pasting bytes, and reads the finished asset URL rather than guessing. Every other client takes the same server through its own config file: see the developer hub.
npx @spicyapi/cli tasks create --model vidu/q4-preview/reference-to-video \
--input-json '{"aspect_ratio":"16:9","duration_seconds":5,"prompt":"A slow aerial shot drifting over a misty pine forest at sunrise","reference_image_urls":["https://example.com/reference.png"],"resolution":"720p"}' \
--waitThe input is this endpoint’s own schema, same as the examples above. --wait polls until the task is finished and prints it; add --yes to skip the confirmation prompt in CI. Set SPICY_API_KEY first — one key works for the API, the CLI and the MCP server.
npx @spicyapi/cli tasks create --model vidu/q4-preview/reference-to-video \
--input-json '{…}' --retention 1h --wait
npx @spicyapi/cli tasks get $TASK_ID # read back the retention that applied--retention only ever shortens the window — 30m, 1h, 7d, or 0 to delete as soon as the task reaches a terminal state. The value that actually applies is the shortest of your request, your account settings and the platform maximum, so read it back from the task record rather than assuming the number you asked for took effect. Passing a longer window than your account allows does nothing.
Install the SpicyAPI nodes from ComfyUI Manager or with the command below, then search for Vidu Q4 Preview · Reference to Video on the canvas. The node has the same options as this page and shows its price before it runs.
cd ComfyUI/custom_nodes
git clone https://github.com/spicyapi-ai/comfyui-spicyapiThe embedded panel reads the same schema as this page, so it cannot drift from the API. If your docs need the content to be searchable, read the JSON under /api/embed/models/ and render it yourself.
<iframe src="https://spicyapi.ai/embed/en/models/vidu-q4-preview/api?model=vidu%2Fq4-preview%2Freference-to-video" width="100%" height="560" loading="lazy" title="Request example" style="border:1px solid rgba(0,0,0,.12)"></iframe>