Yapper API Docs

POST /api/v1/processes with "type": "video-generation". Pick any model — the API does not plan-gate models. GET /api/v1/models returns the same capabilities at runtime.

Seedance generated audio

Seedance 2.x generates a native audio track by default. Set generateAudio: false for a silent video. When referenceAudios are attached, generated audio is required and generateAudio must be true or omitted.

{
  "type": "video-generation",
  "model": "seedance-2.5",
  "input": {
    "prompt": "A silent cinematic landscape reveal",
    "videoLength": 5,
    "generateAudio": false
  }
}

Seedance 2.5 Edit workflow

Use model: "seedance-2.5-edit" for the dedicated base-video workflow. Put the base video first in referenceVideos; later video items are additional references. aspectRatio and durationMode are always "auto", and videoLength is not accepted.

editOperation defaults to "edit". Edit requests require a non-empty prompt; "extend", "sequel", and "prequel" work without one because the server supplies the complete continuation/lead-in instruction.

{
  "type": "video-generation",
  "model": "seedance-2.5-edit",
  "input": {
    "editOperation": "sequel",
    "referenceVideos": [{ "assetId": "base_video_asset_id" }]
  }
}

Input fields

Field	Type	Required	Description
`prompt`	`string`	yes	What to generate.
`editOperation`	`"edit" \| "extend" \| "sequel" \| "prequel"`	no	Only for seedance-2.5-edit. Defaults to "edit"; prompt is optional for "extend", "sequel", and "prequel".
`aspectRatio`	`string`	no	One of the model's aspect ratios, e.g. "16:9". Use "adaptive" when advertised to follow the primary image/video reference.
`resolution`	`number`	no	Output resolution height, e.g. 1080.
`videoLength`	`number`	no	Duration in seconds.
`startingFrameImageUrl`	`string`	no	Image URL used as the first frame. External URLs are auto-imported as assets first.
`endingFrameImageUrl`	`string`	no	Image URL used as the last frame. External URLs are auto-imported as assets first.
`durationMode`	`"fixed" \| "auto"`	no	Use "auto" to match the duration of a supported reference video; defaults to "fixed".
`generateAudio`	`boolean`	no	Whether Seedance should generate a native audio track. Defaults to true. Must remain true when referenceAudios are attached.
`bitrateMode`	`"standard" \| "high"`	no	Seedance 2.x output bitrate preset. Use "high" for a crisper, larger file at the same credit cost; defaults to "standard".
`keyframes`	`{timeSeconds, imageUrl}[]`	no	Timestamped image frames. timeSeconds is measured from the beginning of the generated clip.
`referenceImages`	`({assetId} \| {url})[]`	no	Image references. Send either {assetId} for a Yapper asset or {url} for a public URL to import. Legacy {docId, url} objects remain accepted.
`referenceVideos`	`({assetId} \| {url})[]`	no	Video references. Send either {assetId} for a Yapper asset or {url} for a public URL to import. Legacy {docId, url} objects remain accepted.
`referenceAudios`	`({assetId} \| {url})[]`	no	Audio references. Send either {assetId} for a Yapper asset or {url} for a public URL to import. Legacy {docId, url} objects remain accepted; private storage metadata is resolved server-side.
`disableAutoRetries`	`boolean`	no	Disable automatic same-model retries. Defaults to false, so automatic retries are enabled.
`endingFrameImageRef`	`object`	no	—
`startingFrameImageRef`	`object`	no	—

Referencing media: referenceImageUrls takes plain https image URLs. The normalized referenceImages, referenceVideos, and referenceAudios fields accept either {assetId} for an existing Yapper asset or {url} for public media, which is imported automatically. Provider-native asset:// URIs are internal and are not accepted here. Legacy {docId, url} objects remain accepted for backwards compatibility.

Example

Add "dryRun": true to get the exact credit cost without starting, and always send an Idempotency-Key header on real starts.

POST /api/v1/processes
{
  "type": "video-generation",
  "model": "grok-imagine-v1.5",
  "input": {
    "prompt": "A slow dolly shot over a foggy pine forest at sunrise",
    "videoLength": 1,
    "aspectRatio": "16:9"
  }
}

Models at a glance

Credits are per generated output; ranges span the supported resolutions/lengths (exact prices in each model's section).

Model	Company	Credits	Lengths (s)	Max ref images	Frames	Audio
`wan-3.0-prime`	alibaba	50–750	2–30	10	start + end	yes
`wan-3.0`	alibaba	100–1460	2–30	10	start + end	yes
`flux-3`	black-forest-labs	310–1240	5–20	0	start + end	yes
`minimax-h3`	minimax	70–220	5–15	9	start + end	yes
`minimax-h3-max`	minimax	22–66	5–15	—	start + end	yes
`seedance-2.5`	bytedance	380–2800	4–30	30	start + end	yes
`seedance-2.0`	bytedance	220–840	4–15	9	start + end	yes
`grok-imagine`	xai	58–135	4, 6, 8, 10, 12, 14, 15	—	start	yes
`grok-imagine-v1.5`	xai	20–260	1–15	—	start	yes
`happy-horse`	alibaba	90–450	3–15	5	start	yes
`seedance-2.0-fast`	bytedance	180–680	4–15	9	start + end	yes
`seedance-2.0-mini`	bytedance	40–120	4–15	9	start + end	yes
`wan-2.7`	alibaba	30–150	2, 4, 6, 8, 10	5	start + end	yes
`pixverse-v6`	pixverse	20–370	1, 2, 4, 6, 8, 10, 12, 14, 15	—	start	yes
`kling-3.0`	kuaishou	72–270	4, 6, 8, 10, 12, 14, 15	3	start + end	yes
`kling-3.0-pro`	kuaishou	96–360	4, 6, 8, 10, 12, 14, 15	3	start + end	yes
`gemini-omni-flash`	google	13–397	3–10	10	start + end	yes
`veo3-quality`	google	216	8	3	start + end	yes
`veo3-fast`	google	80	8	3	start + end	yes
`sora-2`	openai	48–240	4, 8, 12, 16, 20	—	start	yes
`sora-2-pro`	openai	430–2150	4, 8, 12, 16, 20	—	start	yes
`seedance-2.0-open`	bytedance	310–1180	4–15	9	start + end	yes
`seedance-2.5-edit`	bytedance	dryRun	—	30	—	yes

WAN 3.0 Prime (wan-3.0-prime)

Alibaba's high-speed WAN 3.0 tier, around 7x faster with the same multimodal inputs and native audio.


Credits	2s: 50 · 3s: 80 · 4s: 100 · 5s: 130 · 6s: 150 · 7s: 180 · 8s: 200 · 9s: 230 · 10s: 250 · 11s: 280 · 12s: 300 · 13s: 330 · 14s: 350 · 15s: 380 · 16s: 400 · 17s: 430 · 18s: 450 · 19s: 470 · 20s: 500 · 21s: 520 · 22s: 550 · 23s: 570 · 24s: 600 · 25s: 620 · 26s: 650 · 27s: 670 · 28s: 700 · 29s: 720 · 30s: 750 (at 1080p)
Pricing	Limited-time launch price; the future standard credit price is 2× the current rate
Aspect ratios	16:9, 4:3, 1:1, 3:4, 9:16
Lengths (s)	2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30
Resolutions	480, 720, 1080
Reference images	up to 10
Reference videos	up to 5 (15s total)
Reference audio	up to 5 (15s total)
Text-only generation	supported
Start/end frames	start + end
Audio output	yes
Max prompt length	5000

WAN 3.0 (wan-3.0)

Alibaba's newest multimodal video model with native audio, start/end frames, and image, video, and audio references.


Credits	2s: 100 · 3s: 150 · 4s: 200 · 5s: 250 · 6s: 300 · 7s: 350 · 8s: 390 · 9s: 440 · 10s: 490 · 11s: 540 · 12s: 590 · 13s: 640 · 14s: 690 · 15s: 730 · 16s: 780 · 17s: 830 · 18s: 880 · 19s: 930 · 20s: 980 · 21s: 1030 · 22s: 1070 · 23s: 1120 · 24s: 1170 · 25s: 1220 · 26s: 1270 · 27s: 1320 · 28s: 1370 · 29s: 1410 · 30s: 1460 (at 1080p)
Aspect ratios	16:9, 4:3, 1:1, 3:4, 9:16
Lengths (s)	2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30
Resolutions	480, 720, 1080
Reference images	up to 10
Reference videos	up to 5 (15s total)
Reference audio	up to 5 (15s total)
Text-only generation	supported
Start/end frames	start + end
Audio output	yes
Max prompt length	5000

FLUX.3 (flux-3)

Black Forest Labs' frontier video model for text, timestamped image keyframes, start/end-frame interpolation, or video continuation, with native audio and clips up to 20 seconds.


Credits	5s: 310 · 6s: 370 · 7s: 440 · 8s: 500 · 9s: 560 · 10s: 620 · 11s: 680 · 12s: 750 · 13s: 810 · 14s: 870 · 15s: 930 · 16s: 990 · 17s: 1060 · 18s: 1120 · 19s: 1180 · 20s: 1240 (at 1080p)
Aspect ratios	16:9, 9:16, 1:1, 21:9, 2:1, 4:3, 3:4, auto (default auto)
Lengths (s)	5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20
Resolutions	720, 1080
Reference images	up to 0
Reference videos	up to 1
Text-only generation	supported
Start/end frames	start + end
Audio output	yes

MiniMax H3 (minimax-h3)

MiniMax's frontier 2K video model for text generation and multimodal image, video, and audio references.


Credits	5s: 70 · 6s: 90 · 7s: 100 · 8s: 120 · 9s: 130 · 10s: 150 · 11s: 160 · 12s: 170 · 13s: 190 · 14s: 200 · 15s: 220 (at 768p)
Aspect ratios	16:9, 21:9, 4:3, 1:1, 3:4, 9:16, adaptive
Lengths (s)	5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
Resolutions	768, 1440
Reference images	up to 9
Reference videos	up to 3 (15s total)
Reference audio	up to 3 (15s total)
Text-only generation	supported
Start/end frames	start + end
Audio output	yes
Max prompt length	2000

MiniMax H3 Max (minimax-h3-max)

Fal's post-trained H3 variant for exceptionally fast 480p and 768p text-to-video or first/end-frame animation with native audio.


Credits	5s: 22 · 6s: 26 · 7s: 31 · 8s: 35 · 9s: 40 · 10s: 44 · 11s: 49 · 12s: 53 · 13s: 57 · 14s: 62 · 15s: 66 (at 480p)
Pricing	Limited-time launch price; the future standard credit price is 2× the current rate
Aspect ratios	16:9, 21:9, 4:3, 1:1, 3:4, 9:16 — startingFrameImageUrl requires "auto" (or omission)
Lengths (s)	5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
Resolutions	480, 768
Text-only generation	supported
Start/end frames	start + end
Audio output	yes

Seedance 2.5 (seedance-2.5)

ByteDance's latest multimodal video model with native audio, 30-second generations, and expanded image, video, and audio references.


Credits	4s: 380 · 5s: 470 · 6s: 570 · 7s: 660 · 8s: 750 · 9s: 850 · 10s: 940 · 11s: 1030 · 12s: 1130 · 13s: 1220 · 14s: 1310 · 15s: 1400 · 16s: 1500 · 17s: 1590 · 18s: 1680 · 19s: 1780 · 20s: 1870 · 21s: 1960 · 22s: 2060 · 23s: 2150 · 24s: 2240 · 25s: 2340 · 26s: 2430 · 27s: 2520 · 28s: 2610 · 29s: 2710 · 30s: 2800 (at 1080p)
Pricing	Limited-time launch price; the future standard credit price is 1.5× the current rate
Aspect ratios	16:9, 4:3, 1:1, 3:4, 9:16, 21:9, adaptive
Lengths (s)	4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30
Resolutions	480, 720, 1080
Bitrate modes	standard, high (default standard)
Generated audio	configurable (default on; required with reference audio)
Reference images	up to 30
Reference videos	up to 10 (30s total)
Reference audio	up to 10 (30s total)
Text-only generation	supported
Start/end frames	start + end
Audio output	yes
Max prompt length	20000

Seedance 2.0 (seedance-2.0)

ByteDance's video model with multi-modal input, lip sync, and multi-shot narrative.


Credits	4s: 220 · 5s: 280 · 6s: 340 · 7s: 390 · 8s: 450 · 9s: 510 · 10s: 560 · 11s: 620 · 12s: 670 · 13s: 730 · 14s: 790 · 15s: 840 (at 1080p)
Aspect ratios	16:9, 4:3, 1:1, 3:4, 9:16, 21:9, adaptive
Lengths (s)	4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
Resolutions	480, 720, 1080, 2160
Bitrate modes	standard, high (default standard)
Generated audio	configurable (default on; required with reference audio)
Reference images	up to 9
Reference videos	up to 3 (15s total)
Reference audio	up to 3 (15s total)
Text-only generation	not supported
Start/end frames	start + end
Audio output	yes
Max prompt length	20000

Grok Imagine (grok-imagine)

Grok's best and latest video model with sound included.


Credits	4s: 58 · 6s: 72 · 8s: 86 · 10s: 100 · 12s: 114 · 14s: 128 · 15s: 135 (at 720p)
Aspect ratios	16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 1:1 — "auto" also accepted with startingFrameImageUrl
Lengths (s)	4, 6, 8, 10, 12, 14, 15
Resolutions	720
Text-only generation	supported
Start/end frames	start
Audio output	yes
Max prompt length	4080

Grok Imagine 1.5 (grok-imagine-v1.5)

xAI's Grok Imagine 1.5 image-to-video model with audio included.


Credits	1s: 20 · 2s: 40 · 3s: 50 · 4s: 70 · 5s: 90 · 6s: 110 · 7s: 120 · 8s: 140 · 9s: 160 · 10s: 170 · 11s: 190 · 12s: 210 · 13s: 230 · 14s: 240 · 15s: 260 (at 480p)
Aspect ratios	16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 1:1 — "auto" also accepted with startingFrameImageUrl
Lengths (s)	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
Resolutions	480, 720
Text-only generation	not supported
Start/end frames	start
Audio output	yes
Max prompt length	4096

Happy Horse (happy-horse)

Alibaba's reference-to-video model with native audio and 720p output.


Credits	3s: 90 · 4s: 120 · 5s: 150 · 6s: 180 · 7s: 210 · 8s: 240 · 9s: 270 · 10s: 300 · 11s: 330 · 12s: 360 · 13s: 390 · 14s: 420 · 15s: 450 (at 720p)
Aspect ratios	16:9, 9:16, 1:1, 4:3, 3:4
Lengths (s)	3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
Resolutions	720
Reference images	up to 5
Text-only generation	not supported
Start/end frames	start
Audio output	yes
Max prompt length	5000

Seedance 2.0 Fast (seedance-2.0-fast)

Fast variant of ByteDance's video model with multi-modal input, lip sync, and multi-shot narrative.


Credits	4s: 180 · 5s: 230 · 6s: 270 · 7s: 320 · 8s: 360 · 9s: 410 · 10s: 450 · 11s: 500 · 12s: 540 · 13s: 590 · 14s: 640 · 15s: 680 (at 1080p)
Aspect ratios	16:9, 4:3, 1:1, 3:4, 9:16, 21:9, adaptive
Lengths (s)	4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
Resolutions	480, 720, 1080, 2160
Bitrate modes	standard, high (default standard)
Generated audio	configurable (default on; required with reference audio)
Reference images	up to 9
Reference videos	up to 3 (15s total)
Reference audio	up to 3 (15s total)
Text-only generation	not supported
Start/end frames	start + end
Audio output	yes
Max prompt length	20000

Seedance 2.0 Mini (seedance-2.0-mini)

ByteDance's economical video model for fast multimodal drafts with native audio.


Credits	4s: 40 · 5s: 40 · 6s: 50 · 7s: 60 · 8s: 60 · 9s: 70 · 10s: 80 · 11s: 90 · 12s: 90 · 13s: 100 · 14s: 110 · 15s: 120 (at 480p)
Aspect ratios	16:9, 4:3, 1:1, 3:4, 9:16, 21:9, adaptive
Lengths (s)	4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
Resolutions	480, 720
Bitrate modes	standard, high (default standard)
Generated audio	configurable (default on; required with reference audio)
Reference images	up to 9
Reference videos	up to 3 (15s total)
Reference audio	up to 3 (15s total)
Text-only generation	not supported
Start/end frames	start + end
Audio output	yes
Max prompt length	20000

WAN 2.7 (wan-2.7)

Wan's latest reference-driven video model with image and video guidance for better consistency.


Credits	2s: 30 · 4s: 60 · 6s: 90 · 8s: 120 · 10s: 150 (at 1080p)
Aspect ratios	16:9, 9:16, 1:1, 4:3, 3:4
Lengths (s)	2, 4, 6, 8, 10
Resolutions	720, 1080
Reference images	up to 5
Reference videos	up to 1
Text-only generation	supported
Start/end frames	start + end
Audio output	yes
Max prompt length	5000

PixVerse V6 (pixverse-v6)

PixVerse's latest Fal-hosted video model with optional audio and first-frame image support.


Credits	1s: 20 · 2s: 50 · 4s: 100 · 6s: 150 · 8s: 200 · 10s: 250 · 12s: 300 · 14s: 350 · 15s: 370 (at 1080p)
Aspect ratios	16:9, 4:3, 1:1, 3:4, 9:16, 2:3, 3:2, 21:9 — startingFrameImageUrl requires "auto" (or omission)
Lengths (s)	1, 2, 4, 6, 8, 10, 12, 14, 15
Resolutions	360, 540, 720, 1080
Text-only generation	supported
Start/end frames	start
Audio output	yes
Max prompt length	2024

Kling 3.0 (kling-3.0)

Video model with start/end frames, references, and native audio.


Credits	4s: 72 · 6s: 108 · 8s: 144 · 10s: 180 · 12s: 216 · 14s: 252 · 15s: 270
Aspect ratios	16:9, 9:16, 1:1 — startingFrameImageUrl and referenceVideos require "auto" (or omission)
Lengths (s)	4, 6, 8, 10, 12, 14, 15
Reference images	up to 3
Reference videos	up to 1
Start/end frames	start + end
Audio output	yes
Max prompt length	2500

Kling 3.0 Pro (kling-3.0-pro)

Higher-cost Kling model with start/end frame, reference, and native audio support.


Credits	4s: 96 · 6s: 144 · 8s: 192 · 10s: 240 · 12s: 288 · 14s: 336 · 15s: 360
Aspect ratios	16:9, 9:16, 1:1 — startingFrameImageUrl and referenceVideos require "auto" (or omission)
Lengths (s)	4, 6, 8, 10, 12, 14, 15
Reference images	up to 3
Reference videos	up to 1
Start/end frames	start + end
Audio output	yes
Max prompt length	2500

Gemini Omni 1.1 Flash (gemini-omni-flash)

Google's fast multimodal video model with native audio, first/end-frame interpolation, video editing and extension, and output up to 4K.


Pricing	Limited-time launch price targeting a -15% margin; the future standard credit price is 1.9× the current rate
Input pricing	The displayed total adds Gemini's billed tokens for every image and the measured input-video duration. An input video without duration metadata is conservatively priced as 10s.
Credits (360p, no media inputs)	3s: 13 · 4s: 18 · 5s: 22 · 6s: 26 · 7s: 31 · 8s: 35 · 9s: 40 · 10s: 44
Credits (720p, no media inputs)	3s: 40 · 4s: 53 · 5s: 66 · 6s: 79 · 7s: 93 · 8s: 106 · 9s: 119 · 10s: 132
Credits (1080p, no media inputs)	3s: 59 · 4s: 79 · 5s: 99 · 6s: 119 · 7s: 139 · 8s: 159 · 9s: 178 · 10s: 198
Credits (4K, no media inputs)	3s: 119 · 4s: 159 · 5s: 198 · 6s: 238 · 7s: 278 · 8s: 317 · 9s: 357 · 10s: 397
Aspect ratios	16:9, 9:16
Resolutions	360p, 720p (default), 1080p (upscaled), 4K (upscaled)
Lengths (s)	3, 4, 5, 6, 7, 8, 9, 10
Reference images	up to 10 total image inputs, including start/end frames
Reference videos	up to 1
Start/end frames	start + end
Text-only generation	supported
Audio output	yes

Veo3.1 Quality (veo3-quality)

Google's highest quality video model with sound included. Blocks celebrities.


Credits	8s: 216
Aspect ratios	16:9, 9:16
Lengths (s)	8
Reference images	up to 3
Text-only generation	supported
Start/end frames	start + end
Audio output	yes

Veo3.1 (veo3-fast)

Google's latest video model with sound included. Blocks celebrities.


Credits	8s: 80
Aspect ratios	16:9, 9:16
Lengths (s)	8
Reference images	up to 3
Text-only generation	supported
Start/end frames	start + end
Audio output	yes

Sora 2 (sora-2)

OpenAI's new bleeding-edge video model with sound included.


Credits	4s: 48 · 8s: 96 · 12s: 144 · 16s: 192 · 20s: 240
Aspect ratios	16:9, 9:16
Lengths (s)	4, 8, 12, 16, 20
Text-only generation	supported
Start/end frames	start
Audio output	yes
Max prompt length	5000

Sora 2 Pro (sora-2-pro)