Kandinsky 6 API Pro & Lite, Kandinsky 6 VSR API, Kandinsky 5 Pro API: Pricing, Documentation

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/kandinsky-6-lite/v1/text-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"prompt": "A paper boat drifting down a rain-soaked city street at night, neon reflections rippling on the water, low tracking shot. Rain patters on the pavement and distant traffic hums.", "aspect_ratio": "16:9", "num_inference_steps": 10, "upscale_factor": 2.25}'

Callback Payload (success)

{
  "request_id": "kandinsky-6-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-lite",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Failure callback shape

{
  "request_id": "kandinsky-6-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "kandinsky-6-lite",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request returns one video. batch_size, num_images, number_of_images, n and count are rejected — send one request per video.
• Every clip is about 5 seconds at 480p: there is no length or resolution setting, and duration, seconds, num_frames, frames_per_second, resolution, width, height and music_length_ms are rejected. For a larger video set upscale_factor; 2.25 turns a 16:9 clip into 1080p.
• Send every field at the top level of the JSON body, with the text in prompt; parameters, content and text are rejected.
• To animate an image, use the Image to Video operation; image_url, image_urls and image are rejected here.
• Use the JSON types shown: num_inference_steps is a whole number (10, not "10"), and generate_audio and enable_prompt_expansion are true or false, not strings.
• When you leave a field out, the defaults apply: 10 steps, no upscaling (480p), 16:9 and a synchronized audio track.
• Only upscale_factor and seed accept null. Omit the other optional fields you do not use — do not send them as null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url and webhook_mode in the JSON body are rejected.

Length, Resolution and Upscaling

Every request renders one clip of about 5 seconds at 480p. There is no length or resolution parameter: to get a larger video, set upscale_factor, and the clip is upscaled with Kandinsky 6 video super-resolution before it is returned.

Each upscale factor costs more than no upscaling, and 4 costs the most; see How billing is measured below.

Aspect ratios

16:9, 9:16, 1:1, 4:3, 3:4. 16:9 is the default; landscape, portrait and square shapes are all available.

Inference steps

num_inference_steps sets the number of denoising steps, from 4 to 16. Lite is tuned for 10, the default. Fewer steps cost less and more steps cost more.

Example: no audio track

{
  "prompt": "Time-lapse of clouds rolling over a mountain ridge at sunset, wide shot.",
  "generate_audio": false
}

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing prompt (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' is required",
  "violations": [
    "'prompt' is required"
  ]
}

A value that is not allowed (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'aspect_ratio' must be one of: \"16:9\", \"9:16\", \"1:1\", \"4:3\", \"3:4\" (got \"21:9\")",
  "violations": [
    "'aspect_ratio' must be one of: \"16:9\", \"9:16\", \"1:1\", \"4:3\", \"3:4\" (got \"21:9\")"
  ]
}

A value out of range (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'num_inference_steps' must be <= 16 (got 20)",
  "violations": [
    "'num_inference_steps' must be <= 16 (got 20)"
  ]
}

A number sent as a string (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'num_inference_steps' must be integer (got string)",
  "violations": [
    "'num_inference_steps' must be integer (got string)"
  ]
}

A field this operation does not accept (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'duration' must be one of:  (got 10); 'duration' is not supported: Kandinsky 6.0 Lite always renders a 5-second 480p clip and has no length setting. For a larger video set upscale_factor (2.25 gives 1080p; 2 and 4 are also accepted).",
  "violations": [
    "'duration' must be one of:  (got 10)",
    "'duration' is not supported: Kandinsky 6.0 Lite always renders a 5-second 480p clip and has no length setting. For a larger video set upscale_factor (2.25 gives 1080p; 2 and 4 are also accepted)."
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then download the video from output.media_url. Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/kandinsky-6-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "kandinsky-6-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-lite",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/kandinsky-6-lite/v1/text-to-video.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, download the video from output.media_url.

How billing is measured

Billing is measured per video, by the number of inference steps and the upscale factor: the amount for this operation is set by num_inference_steps (10 when you leave it out — more steps cost more) and by upscale_factor (no upscaling when you leave it out — each factor costs more than none, and 4 costs the most). Each request generates one clip of about 5 seconds; the prompt, aspect ratio, audio, prompt expansion and seed do not change the amount. The amount is placed as a hold when the request is submitted and released in full if the request fails.

The current rate is shown in the pricing panel on this page.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/kandinsky-6-lite/v1/image-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"prompt": "Slow push-in as the subject turns toward the camera, hair moving in a light breeze, warm natural light. Soft wind and distant birdsong.", "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg", "num_inference_steps": 10, "upscale_factor": 2.25}'

Callback Payload (success)

{
  "request_id": "kandinsky-6-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-lite",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Failure callback shape

{
  "request_id": "kandinsky-6-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "kandinsky-6-lite",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request returns one video. batch_size, num_images, number_of_images, n and count are rejected — send one request per video.
• Every clip is about 5 seconds at 480p: there is no length or resolution setting, and duration, seconds, num_frames, frames_per_second, resolution, width, height and music_length_ms are rejected. For a larger video set upscale_factor; 2.25 turns a 16:9 clip into 1080p.
• Send every field at the top level of the JSON body, with the text in prompt; parameters, content and text are rejected.
• Send exactly one image as image_url; image_urls and image are rejected.
• Use the JSON types shown: num_inference_steps is a whole number (10, not "10"), and generate_audio and enable_prompt_expansion are true or false, not strings.
• When you leave a field out, the defaults apply: 10 steps, no upscaling (480p), auto aspect ratio and a synchronized audio track.
• Only upscale_factor and seed accept null. Omit the other optional fields you do not use — do not send them as null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url and webhook_mode in the JSON body are rejected.

Length, Resolution and Upscaling

Every request renders one clip of about 5 seconds at 480p. There is no length or resolution parameter: to get a larger video, set upscale_factor, and the clip is upscaled with Kandinsky 6 video super-resolution before it is returned.

Each upscale factor costs more than no upscaling, and 4 costs the most; see How billing is measured below.

Shape

With the default aspect_ratio of auto, the supported ratio closest to your image's is used and the image is center-cropped to it. Set 16:9, 9:16, 1:1, 4:3 or 3:4 to choose the shape yourself.

Inference steps

num_inference_steps sets the number of denoising steps, from 4 to 16. Lite is tuned for 10, the default. Fewer steps cost less and more steps cost more.

Source Image

• Send exactly one image in image_url: a publicly reachable https URL, or a data URI such as data:image/png;base64,.... A list of images is rejected.
• Use a JPEG, PNG or WebP file. The image is downloaded when you submit: a URL that cannot be downloaded, or that returns a web page or an empty file instead of an image, is refused at once with a 400.
• The image becomes the first frame of the video. With the default aspect_ratio of auto, the supported ratio closest to your image's is used and the image is center-cropped to it; set a ratio to choose the shape yourself.
• Describe the motion and the sound in prompt.

Example: send the image as a data URI, in portrait

{
  "prompt": "Slow push-in as the subject turns toward the light. Soft room tone.",
  "image_url": "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQ...",
  "aspect_ratio": "9:16"
}

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. An image that cannot be downloaded is also refused at submit with a 400, and its hold is released at once. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing prompt (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' is required",
  "violations": [
    "'prompt' is required"
  ]
}

A value that is not allowed (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'aspect_ratio' must be one of: \"16:9\", \"9:16\", \"1:1\", \"4:3\", \"3:4\", \"auto\" (got \"21:9\")",
  "violations": [
    "'aspect_ratio' must be one of: \"16:9\", \"9:16\", \"1:1\", \"4:3\", \"3:4\", \"auto\" (got \"21:9\")"
  ]
}

A value out of range (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'num_inference_steps' must be <= 16 (got 20)",
  "violations": [
    "'num_inference_steps' must be <= 16 (got 20)"
  ]
}

A number sent as a string (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'num_inference_steps' must be integer (got string)",
  "violations": [
    "'num_inference_steps' must be integer (got string)"
  ]
}

A field this operation does not accept (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'duration' must be one of:  (got 10); 'duration' is not supported: Kandinsky 6.0 Lite always renders a 5-second 480p clip and has no length setting. For a larger video set upscale_factor (2.25 gives 1080p; 2 and 4 are also accepted).",
  "violations": [
    "'duration' must be one of:  (got 10)",
    "'duration' is not supported: Kandinsky 6.0 Lite always renders a 5-second 480p clip and has no length setting. For a larger video set upscale_factor (2.25 gives 1080p; 2 and 4 are also accepted)."
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then download the video from output.media_url. Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/kandinsky-6-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "kandinsky-6-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-lite",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/kandinsky-6-lite/v1/image-to-video.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, download the video from output.media_url.

How billing is measured

Billing is measured per video, by the number of inference steps and the upscale factor: the amount for this operation is set by num_inference_steps (10 when you leave it out — more steps cost more) and by upscale_factor (no upscaling when you leave it out — each factor costs more than none, and 4 costs the most). Each request generates one clip of about 5 seconds; the prompt, aspect ratio, audio, prompt expansion and seed do not change the amount. The amount is placed as a hold when the request is submitted and released in full if the request fails.

The current rate is shown in the pricing panel on this page.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/kandinsky-6-pro/v1/text-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"prompt": "A red fox trotting through fresh snow at dawn in a pine forest, its breath visible in the cold air, slow tracking shot. Snow crunches under its paws.", "negative_prompt": "blurry, low quality, distorted faces, text, watermark", "aspect_ratio": "16:9", "num_inference_steps": 50, "upscale_factor": 2.25}'

Callback Payload (success)

{
  "request_id": "kandinsky-6-pro_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-pro",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Failure callback shape

{
  "request_id": "kandinsky-6-pro_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "kandinsky-6-pro",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request returns one video. batch_size, num_images, number_of_images, n and count are rejected — send one request per video.
• Every clip is about 5 seconds at 480p: there is no length or resolution setting, and duration, seconds, num_frames, frames_per_second, resolution, width, height and music_length_ms are rejected. For a larger video set upscale_factor; 2.25 turns a 16:9 clip into 1080p.
• Send every field at the top level of the JSON body, with the text in prompt; parameters, content and text are rejected.
• To animate an image, use the Image to Video operation; image_url, image_urls and image are rejected here.
• Use the JSON types shown: num_inference_steps is a whole number (50, not "50"), and generate_audio and enable_prompt_expansion are true or false, not strings.
• When you leave a field out, the defaults apply: 50 steps, no upscaling (480p), 16:9 and a synchronized audio track.
• Only upscale_factor, seed and negative_prompt accept null. Omit the other optional fields you do not use — do not send them as null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url and webhook_mode in the JSON body are rejected.

Length, Resolution and Upscaling

Every request renders one clip of about 5 seconds at 480p. There is no length or resolution parameter: to get a larger video, set upscale_factor, and the clip is upscaled with Kandinsky 6 video super-resolution before it is returned.

Each upscale factor costs more than no upscaling, and 4 costs the most; see How billing is measured below.

Aspect ratios

16:9, 9:16, 1:1, 4:3, 3:4. 16:9 is the default; landscape, portrait and square shapes are all available.

Inference steps

num_inference_steps sets the number of denoising steps, from 8 to 50. 50 is the default and the most; fewer steps cost less.

Example: no audio track

{
  "prompt": "Time-lapse of clouds rolling over a mountain ridge at sunset, wide shot.",
  "generate_audio": false
}

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing prompt (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' is required",
  "violations": [
    "'prompt' is required"
  ]
}

A value that is not allowed (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'aspect_ratio' must be one of: \"16:9\", \"9:16\", \"1:1\", \"4:3\", \"3:4\" (got \"21:9\")",
  "violations": [
    "'aspect_ratio' must be one of: \"16:9\", \"9:16\", \"1:1\", \"4:3\", \"3:4\" (got \"21:9\")"
  ]
}

A value out of range (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'num_inference_steps' must be <= 50 (got 54)",
  "violations": [
    "'num_inference_steps' must be <= 50 (got 54)"
  ]
}

A number sent as a string (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'num_inference_steps' must be integer (got string)",
  "violations": [
    "'num_inference_steps' must be integer (got string)"
  ]
}

A field this operation does not accept (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'duration' must be one of:  (got 10); 'duration' is not supported: Kandinsky 6.0 Pro always renders a 5-second 480p clip and has no length setting. For a larger video set upscale_factor (2.25 gives 1080p; 2 and 4 are also accepted).",
  "violations": [
    "'duration' must be one of:  (got 10)",
    "'duration' is not supported: Kandinsky 6.0 Pro always renders a 5-second 480p clip and has no length setting. For a larger video set upscale_factor (2.25 gives 1080p; 2 and 4 are also accepted)."
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then download the video from output.media_url. Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/kandinsky-6-pro_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "kandinsky-6-pro_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-pro",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/kandinsky-6-pro/v1/text-to-video.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, download the video from output.media_url.

How billing is measured

Billing is measured per video, by the number of inference steps and the upscale factor: the amount for this operation is set by num_inference_steps (50 when you leave it out — more steps cost more) and by upscale_factor (no upscaling when you leave it out — each factor costs more than none, and 4 costs the most). Each request generates one clip of about 5 seconds; the prompt, aspect ratio, audio, prompt expansion, guidance scale, negative prompt and seed do not change the amount. The amount is placed as a hold when the request is submitted and released in full if the request fails.

The current rate is shown in the pricing panel on this page.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/kandinsky-6-pro/v1/image-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"prompt": "The subject smiles and looks out of frame as the camera slowly orbits, golden hour light, shallow depth of field. Gentle ambient street sounds.", "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg", "num_inference_steps": 50, "upscale_factor": 2.25}'

Callback Payload (success)

{
  "request_id": "kandinsky-6-pro_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-pro",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Failure callback shape

{
  "request_id": "kandinsky-6-pro_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "kandinsky-6-pro",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request returns one video. batch_size, num_images, number_of_images, n and count are rejected — send one request per video.
• Every clip is about 5 seconds at 480p: there is no length or resolution setting, and duration, seconds, num_frames, frames_per_second, resolution, width, height and music_length_ms are rejected. For a larger video set upscale_factor; 2.25 turns a 16:9 clip into 1080p.
• Send every field at the top level of the JSON body, with the text in prompt; parameters, content and text are rejected.
• Send exactly one image as image_url; image_urls and image are rejected.
• Use the JSON types shown: num_inference_steps is a whole number (50, not "50"), and generate_audio and enable_prompt_expansion are true or false, not strings.
• When you leave a field out, the defaults apply: 50 steps, no upscaling (480p), auto aspect ratio and a synchronized audio track.
• Only upscale_factor, seed and negative_prompt accept null. Omit the other optional fields you do not use — do not send them as null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url and webhook_mode in the JSON body are rejected.

Length, Resolution and Upscaling

Every request renders one clip of about 5 seconds at 480p. There is no length or resolution parameter: to get a larger video, set upscale_factor, and the clip is upscaled with Kandinsky 6 video super-resolution before it is returned.

Each upscale factor costs more than no upscaling, and 4 costs the most; see How billing is measured below.

Shape

With the default aspect_ratio of auto, the supported ratio closest to your image's is used and the image is center-cropped to it. Set 16:9, 9:16, 1:1, 4:3 or 3:4 to choose the shape yourself.

Inference steps

num_inference_steps sets the number of denoising steps, from 8 to 50. 50 is the default and the most; fewer steps cost less.

Source Image

• Send exactly one image in image_url: a publicly reachable https URL, or a data URI such as data:image/png;base64,.... A list of images is rejected.
• Use a JPEG, PNG or WebP file. The image is downloaded when you submit: a URL that cannot be downloaded, or that returns a web page or an empty file instead of an image, is refused at once with a 400.
• The image becomes the first frame of the video. With the default aspect_ratio of auto, the supported ratio closest to your image's is used and the image is center-cropped to it; set a ratio to choose the shape yourself.
• Describe the motion and the sound in prompt.

Example: send the image as a data URI, in portrait

{
  "prompt": "Slow push-in as the subject turns toward the light. Soft room tone.",
  "image_url": "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQ...",
  "aspect_ratio": "9:16"
}

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. An image that cannot be downloaded is also refused at submit with a 400, and its hold is released at once. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing prompt (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' is required",
  "violations": [
    "'prompt' is required"
  ]
}

A value that is not allowed (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'aspect_ratio' must be one of: \"16:9\", \"9:16\", \"1:1\", \"4:3\", \"3:4\", \"auto\" (got \"21:9\")",
  "violations": [
    "'aspect_ratio' must be one of: \"16:9\", \"9:16\", \"1:1\", \"4:3\", \"3:4\", \"auto\" (got \"21:9\")"
  ]
}

A value out of range (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'num_inference_steps' must be <= 50 (got 54)",
  "violations": [
    "'num_inference_steps' must be <= 50 (got 54)"
  ]
}

A number sent as a string (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'num_inference_steps' must be integer (got string)",
  "violations": [
    "'num_inference_steps' must be integer (got string)"
  ]
}

A field this operation does not accept (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'duration' must be one of:  (got 10); 'duration' is not supported: Kandinsky 6.0 Pro always renders a 5-second 480p clip and has no length setting. For a larger video set upscale_factor (2.25 gives 1080p; 2 and 4 are also accepted).",
  "violations": [
    "'duration' must be one of:  (got 10)",
    "'duration' is not supported: Kandinsky 6.0 Pro always renders a 5-second 480p clip and has no length setting. For a larger video set upscale_factor (2.25 gives 1080p; 2 and 4 are also accepted)."
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then download the video from output.media_url. Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/kandinsky-6-pro_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "kandinsky-6-pro_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-pro",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/kandinsky-6-pro/v1/image-to-video.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, download the video from output.media_url.

How billing is measured

Billing is measured per video, by the number of inference steps and the upscale factor: the amount for this operation is set by num_inference_steps (50 when you leave it out — more steps cost more) and by upscale_factor (no upscaling when you leave it out — each factor costs more than none, and 4 costs the most). Each request generates one clip of about 5 seconds; the prompt, aspect ratio, audio, prompt expansion, guidance scale, negative prompt and seed do not change the amount. The amount is placed as a hold when the request is submitted and released in full if the request fails.

The current rate is shown in the pricing panel on this page.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/kandinsky-6-vsr/v1/upscale-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/kandinsky-6-pro_01a11b22-ab6f-79b3-3afb-38fe14d2fbb54/output.mp4", "upscale_factor": 2.25}'

Callback Payload (success)

{
  "request_id": "kandinsky-6-vsr_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-vsr",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Failure callback shape

{
  "request_id": "kandinsky-6-vsr_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "kandinsky-6-vsr",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request upscales one video. batch_size, num_images, number_of_images, n and count are rejected — send one request per video.
• Send the video as video_url: a publicly reachable https link to an MP4 or MOV file; image_url and image_urls are rejected.
• There is no prompt and no step setting: prompt and num_inference_steps are rejected.
• The output size comes only from your video and upscale_factor, and the length from your video: duration, seconds, num_frames, frames_per_second, resolution, width, height and music_length_ms are rejected.
• Send every field at the top level of the JSON body; parameters, content and text are rejected.
• When you leave upscale_factor out, 2.25 applies.
• Only seed accepts null. Omit the other optional fields you do not use — do not send them as null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url and webhook_mode in the JSON body are rejected.

Input Video

• Send exactly one video in video_url: a publicly reachable https link to an MP4 or MOV file. A list of videos is rejected.
• The frame size and length are read from the file header when you submit, because they set the amount billed. Send a direct, publicly readable link to an MP4 or MOV file: a web page, a link that needs a login, cannot be downloaded or answers too slowly, or a WebM, MKV, GIF or other file that is not MP4 or MOV is refused at once with a 400, and nothing is charged.
• Clips are resampled to 24 frames per second and capped at 121 frames, about 5 seconds; send clips of 5 seconds or less.
• The output is your video's width and height multiplied by upscale_factor; see Upscale Factors below.

Example: upscale by 4

{
  "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/kandinsky-6-pro_01a11b22-ab6f-79b3-3afb-38fe14d2fbb54/output.mp4",
  "upscale_factor": 4
}

Upscale Factors

The output is your video's width and height multiplied by upscale_factor.

Because billing counts output pixels, the amount grows with the square of the factor: for the same video, 4 bills four times the pixels of 2.

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. A video whose frame size and length cannot be read is also refused at submit with a 400, before any hold is placed. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing video_url (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'video_url' is required",
  "violations": [
    "'video_url' is required"
  ]
}

A value that is not allowed (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'upscale_factor' must be one of: 2, 2.25, 4 (got 3)",
  "violations": [
    "'upscale_factor' must be one of: 2, 2.25, 4 (got 3)"
  ]
}

A list instead of one URL (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'video_url' must be string (got array)",
  "violations": [
    "'video_url' must be string (got array)"
  ]
}

A field this operation does not accept (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' must be one of:  (got \"Sharpen the details\"); 'prompt' is not accepted: Kandinsky 6.0 VSR upscales the video in 'video_url' and takes no prompt.",
  "violations": [
    "'prompt' must be one of:  (got \"Sharpen the details\")",
    "'prompt' is not accepted: Kandinsky 6.0 VSR upscales the video in 'video_url' and takes no prompt."
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then download the upscaled video from output.media_url. Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/kandinsky-6-vsr_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "kandinsky-6-vsr_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-vsr",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/kandinsky-6-vsr/v1/upscale-video.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, download the upscaled video from output.media_url.

How billing is measured

Billing is measured per output megapixel per frame: the output width times the output height times the number of frames, counted in millions of pixels and rounded up to a whole number. The output size is your video's width and height multiplied by upscale_factor (2.25 when you leave it out), so the amount grows with your video's size and with the square of the factor. The frame count is your video's length at 24 frames per second, rounded up to a whole frame, at most 121 frames (about 5 seconds). Both are read from the video file when you submit, and the amount is placed as a hold then. When the request completes, the charge is settled against the delivered video and can only go down; a failed request is not billed — its hold is released in full.

The current rate is shown in the pricing panel on this page.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/kandinsky-6-vsr-lite/v1/upscale-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/kandinsky-6-pro_01a11b22-ab6f-79b3-3afb-38fe14d2fbb54/output.mp4", "upscale_factor": 2.25}'

Callback Payload (success)

{
  "request_id": "kandinsky-6-vsr-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-vsr-lite",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Failure callback shape

{
  "request_id": "kandinsky-6-vsr-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "kandinsky-6-vsr-lite",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on every poll while the request is processing (status PROCESSING), plus the final COMPLETED/ERROR callback.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Good to know

• Each request upscales one video. batch_size, num_images, number_of_images, n and count are rejected — send one request per video.
• Send the video as video_url: a publicly reachable https link to an MP4 or MOV file; image_url and image_urls are rejected.
• There is no prompt and no step setting: prompt and num_inference_steps are rejected.
• The output size comes only from your video and upscale_factor, and the length from your video: duration, seconds, num_frames, frames_per_second, resolution, width, height and music_length_ms are rejected.
• Send every field at the top level of the JSON body; parameters, content and text are rejected.
• When you leave upscale_factor out, 2.25 applies.
• Only seed accepts null. Omit the other optional fields you do not use — do not send them as null.
• Results are always returned as a URL in output.media_url; sync_mode is accepted only as false.
• Webhooks are set with the X-Webhook-URL / X-Webhook-Mode headers; webhook_url and webhook_mode in the JSON body are rejected.

Input Video

• Send exactly one video in video_url: a publicly reachable https link to an MP4 or MOV file. A list of videos is rejected.
• The frame size and length are read from the file header when you submit, because they set the amount billed. Send a direct, publicly readable link to an MP4 or MOV file: a web page, a link that needs a login, cannot be downloaded or answers too slowly, or a WebM, MKV, GIF or other file that is not MP4 or MOV is refused at once with a 400, and nothing is charged.
• Clips are resampled to 24 frames per second and capped at 121 frames, about 5 seconds; send clips of 5 seconds or less.
• The output is your video's width and height multiplied by upscale_factor; see Upscale Factors below.

Example: upscale by 4

{
  "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/kandinsky-6-pro_01a11b22-ab6f-79b3-3afb-38fe14d2fbb54/output.mp4",
  "upscale_factor": 4
}

Upscale Factors

The output is your video's width and height multiplied by upscale_factor.

Because billing counts output pixels, the amount grows with the square of the factor: for the same video, 4 bills four times the pixels of 2.

Request Headers

Response Handling

Parameter problems are caught when you submit and returned as a 400, before any hold is placed on your balance. A video whose frame size and length cannot be read is also refused at submit with a 400, before any hold is placed. Problems that need the model to run are reported through the status endpoint as status: "ERROR", with a message in error saying what went wrong. Failed requests are not billed — the hold is released in full.

Example error responses

A refused request returns error, a message that joins every problem found, and the same problems one by one in violations.

Missing video_url (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'video_url' is required",
  "violations": [
    "'video_url' is required"
  ]
}

A value that is not allowed (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'upscale_factor' must be one of: 2, 2.25, 4 (got 3)",
  "violations": [
    "'upscale_factor' must be one of: 2, 2.25, 4 (got 3)"
  ]
}

A list instead of one URL (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'video_url' must be string (got array)",
  "violations": [
    "'video_url' must be string (got array)"
  ]
}

A field this operation does not accept (synchronous 400)

{
  "error": "Invalid request parameters",
  "message": "'prompt' must be one of:  (got \"Sharpen the details\"); 'prompt' is not accepted: Kandinsky 6.0 VSR Lite upscales the video in 'video_url' and takes no prompt.",
  "violations": [
    "'prompt' must be one of:  (got \"Sharpen the details\")",
    "'prompt' is not accepted: Kandinsky 6.0 VSR Lite upscales the video in 'video_url' and takes no prompt."
  ]
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED or ERROR, then download the upscaled video from output.media_url. Treat any status other than QUEUED, PROCESSING or COMPLETED as a failure.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

cURL example

curl 'https://gateway.pixazo.ai/v2/requests/status/kandinsky-6-vsr-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "kandinsky-6-vsr-lite_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-6-vsr-lite",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-10-08T10:04:54.387Z",
  "completed_at": "2026-10-08 10:06:21"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or ERROR (failure).

Typical workflow

1. Send the request to POST https://gateway.pixazo.ai/kandinsky-6-vsr-lite/v1/upscale-video.
2. Save the request_id from the 202 response.
3. Poll GET /v2/requests/status/{request_id} every 10 seconds or so, or use a Webhook instead.
4. When status is COMPLETED, download the upscaled video from output.media_url.

How billing is measured

Billing is measured per output megapixel per frame: the output width times the output height times the number of frames, counted in millions of pixels and rounded up to a whole number. The output size is your video's width and height multiplied by upscale_factor (2.25 when you leave it out), so the amount grows with your video's size and with the square of the factor. The frame count is your video's length at 24 frames per second, rounded up to a whole frame, at most 121 frames (about 5 seconds). Both are read from the video file when you submit, and the amount is placed as a hold then. When the request completes, the charge is settled against the delivered video and can only go down; a failed request is not billed — its hold is released in full.

The current rate is shown in the pricing panel on this page.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Add the X-Webhook-URL header to your submit request to receive a POST callback when the job completes — no polling required.

Using curl? These are HTTP request headers — pass each with -H, e.g. -H "X-Webhook-URL: https://your-server.com/webhook/callback". Do not paste them as bare lines, and end every line of a multi-line command with \.

Webhook Headers

Example: enable webhook

X-Webhook-URL: https://your-server.com/webhook/callback
X-Webhook-Mode: terminal

Callback Payload

Your endpoint receives a POST application/json with the same shape as the GET /v2/requests/status/{request_id} response. Example terminal callback (mode terminal):

{
  "request_id": "kandinsky-5-0-pro-953_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-5-0-pro-953",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/kandinsky-5-0-pro-953_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.png"
    ],
    "media_type": "image/png"
  },
  "created_at": "2026-05-22T13:17:32.110Z",
  "updated_at": "2026-05-22 13:19:23",
  "completed_at": "2026-05-22 13:19:23"
}

Failure callback shape

{
  "request_id": "kandinsky-5-0-pro-953_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "kandinsky-5-0-pro-953",
  "error": "Description of the error",
  "output": null,
  "created_at": "...",
  "updated_at": "...",
  "completed_at": "..."
}

Delivery semantics

• terminal mode (default) — exactly one POST when the request reaches a terminal status. No callback during PROCESSING.
• sync mode — POST on every status poll (with delay capped at ~15s) plus a final POST at terminal status. Use when you want progress updates.
• Idempotency — use request_id as your idempotency key. Network retries can deliver the same callback more than once; your handler must tolerate duplicates.
• Response — respond 200 OK within a few seconds. The queue does not block on slow handlers, but persistent failures may stop further deliveries.
• HTTPS required — plain http:// URLs are rejected.

Request Parameters - Kandinsky 5.0 Pro generate request

Content Item Types & Limits

Response

{
  "request_id": "kandinsky-5-0-pro-953_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/kandinsky-5-0-pro-953_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Request Headers

Response Handling

Common status codes for Kandinsky 5.0 Pro generate request.

Error Responses

Queue system errors and model validation errors.

Queue System Errors

// 402 — Insufficient balance
{
  "error": "Insufficient Balance",
  "message": "Your wallet does not have enough balance."
}

// 400 — Model not found
{
  "error": "Model not found",
  "message": "Model 'kandinsky-5-0-pro-953' not found or is disabled"
}

Error via Status/Webhook

{
  "request_id": "kandinsky-5-0-pro-953_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "kandinsky-5-0-pro-953",
  "error": "Description of the error",
  "output": null
}

Retrieving Results

Poll the universal status endpoint to check progress and retrieve results.

Endpoint

GET https://gateway.pixazo.ai/v2/requests/status/{request_id}
Ocp-Apim-Subscription-Key: YOUR_API_KEY

curl -H "Ocp-Apim-Subscription-Key: YOUR_API_KEY" \
  "https://gateway.pixazo.ai/v2/requests/status/kandinsky-5-0-pro-953_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"

{
  "request_id": "kandinsky-5-0-pro-953_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "kandinsky-5-0-pro-953",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/kandinsky-5-0-pro-953_019dxxxx-xxxx/output.ext"
    ],
    "media_type": "application/octet-stream"
  },
  "created_at": "2026-03-31T10:00:00.000Z",
  "updated_at": "2026-03-31T10:00:15.000Z",
  "completed_at": "2026-03-31T10:00:15.000Z"
}

Response Fields

Status Values

Status Flow

QUEUED → PROCESSING → COMPLETED
                    → FAILED
                    → ERROR

Typical Workflow

1. Send a generate request to the API endpoint
2. Save the request_id from the response
3. Poll every 5-10 seconds: GET /v2/requests/status/{request_id}
4. When status is "COMPLETED", download from output.media_url

Tip: Use X-Webhook-URL header to get a callback instead of polling.