Imagine Models

Vofy Model Selection

Choose a model before composing vofy image create or vofy video create. When exact limits matter, inspect the generated references or run vofy models <name>.

Selection Steps

1. Classify the task as image or video, then identify the required mode. 2. Pick from Quick Picks for common requests. 3. Confirm exact allowed aspect ratios, resolutions, durations, source inputs, and pricing in the generated references. 4. Avoid model-specific flags until confirmed by vofy models <name> or the references.

Quick Picks

When To Inspect Details

• User asks for exact cost, ratio, resolution, duration, or output count.
• User supplies source media and the mode may derive output dimensions from it.
• User asks for audio, search grounding, transparent background, multi-shot, motion control, or mixed references.
• A create command fails with an unsupported mode, flag, or value.

Mode abbreviations in references: t2i, i2i, t2v, i2v, interp, ref, mm, v2v, ext, mc.

References

• Use image-models.md for exact image modes, aspect ratios, resolutions, special parameters, and pricing.
• Use video-models.md for exact video modes, durations, source constraints, special parameters, and pricing.
• Use vofy models --type image, vofy models --type video, or vofy models <name> for live CLI output.

Image Models — Detailed Reference

Auto-generated from model-manifest.json. Hand-written notes in NOTES blocks are preserved.

gemini-2.5-flash-image

Per-mode constraints:

• text_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto
• image_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, auto
• image_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, auto
• route_c (1x)
• text_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto
• image_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto

Route-specific restrictions:

• route_a: when resolution=2K/4K — Route A only supports 1K and Auto resolutions

<!-- NOTES:gemini-2.5-flash-image --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

gemini-3-pro-image-preview

Per-mode constraints:

• text_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto
• image_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto

Pricing:

• Base: 8 credits
• resolution=1K/2K: 8 credits
• resolution=4K: 16 credits

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto
• image_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto
• route_c (0.75x)
• text_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto
• image_to_image: AR: 1:1, 2:3, 3:4, 4:5, 9:16, 3:2, 4:3, 5:4, 16:9, 21:9, auto, Res: 1K, 2K, 4K, auto

<!-- NOTES:gemini-3-pro-image-preview --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

gemini-3.1-flash-image-preview

Per-mode constraints:

• text_to_image: AR: 1:1, 1:4, 1:8, 2:3, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 3:2, 16:9, 21:9, auto, Res: 512, 1K, 2K, 4K, auto
• image_to_image: AR: 1:1, 1:4, 1:8, 2:3, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 3:2, 16:9, 21:9, auto, Res: 512, 1K, 2K, 4K, auto

Special Parameters:

• --web-search — Let the model use live web results as extra context when the prompt depends on current facts, brands, places, or public references. Default: false.
• --image-search — Let the model retrieve reference imagery for better visual grounding when the prompt depends on recognizable objects, products, or styles. Default: false.
• --reasoning-effort <minimal|high> — Control how much reasoning the model spends interpreting the prompt before generating. Higher effort can improve accuracy but usually adds latency. Default: minimal.

Pricing:

• Base: 6 credits
• resolution=512: 4 credits
• resolution=1K: 6 credits
• resolution=2K: 10 credits
• resolution=4K: 14 credits
• Surcharge: web_search=true adds +1 credits
• Surcharge: image_search=true adds +1 credits
• Surcharge: reasoning_effort=high adds +1 credits

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 1:4, 1:8, 2:3, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 3:2, 16:9, 21:9, auto, Res: 512, 1K, 2K, 4K, auto
• image_to_image: AR: 1:1, 1:4, 1:8, 2:3, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 3:2, 16:9, 21:9, auto, Res: 512, 1K, 2K, 4K, auto
• route_c (0.5x)
• text_to_image: AR: 1:1, 1:4, 1:8, 2:3, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 3:2, 16:9, 21:9, auto, Res: 512, 1K, 2K, 4K, auto
• image_to_image: AR: 1:1, 1:4, 1:8, 2:3, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 3:2, 16:9, 21:9, auto, Res: 512, 1K, 2K, 4K, auto

<!-- NOTES:gemini-3.1-flash-image-preview --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

gpt-image-2

Per-mode constraints:

• text_to_image: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto
• image_to_image: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto
• inpainting: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto

Special Parameters:

• --quality <low|medium|high|auto> — Choose whether the model prioritizes speed, balances quality automatically, or renders with higher image fidelity. Default: auto.
• --background <auto|opaque> — Choose whether the image background stays opaque or is decided automatically by the model. Default: auto.

Pricing:

• Base: 12 credits
• 12 pricing tiers (varies by resolution/duration/mode)
• Range: 2–48 credits

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto
• image_to_image: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto
• inpainting: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto
• route_c (0.75x)
• text_to_image: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto
• image_to_image: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto
• inpainting: AR: 1:1, 1:2, 2:3, 3:4, 4:5, 5:4, 4:3, 3:2, 2:1, 9:16, 9:19.5, 9:20, 16:9, 19.5:9, 20:9, 21:9, auto, Res: 1K, 2K, 3K, 4K, auto

<!-- NOTES:gpt-image-2 --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

gpt-image-1.5

Per-mode constraints:

• text_to_image: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto
• image_to_image: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto
• inpainting: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto

Special Parameters:

• --quality <low|medium|high|auto> — Choose whether the model prioritizes speed, balances quality automatically, or renders with higher image fidelity. Default: auto.
• --background <transparent|opaque|auto> — Choose whether the image background stays transparent, becomes a solid backdrop, or is decided automatically by the model. Default: auto.

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto
• image_to_image: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto
• inpainting: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto
• route_c (0.5x)
• text_to_image: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto
• image_to_image: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto
• inpainting: AR: 1:1, 2:3, 3:2, auto, Res: 1K, auto

<!-- NOTES:gpt-image-1.5 --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

grok-imagine-image

Per-mode constraints:

• text_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto
• image_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto

Input Constraints:

• image: formats: jpg/jpeg/png

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto
• image_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto

<!-- NOTES:grok-imagine-image --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

grok-imagine-image-quality

Per-mode constraints:

• text_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto
• image_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto

Input Constraints:

• image: formats: jpg/jpeg/png

Pricing:

• Base: 8 credits
• resolution=1K: 6 credits
• resolution=2K/auto: 8 credits

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto
• image_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto

<!-- NOTES:grok-imagine-image-quality --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

grok-imagine-image-pro

Per-mode constraints:

• text_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto
• image_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto

Input Constraints:

• image: formats: jpg/jpeg/png

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto
• image_to_image: AR: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, 9:20, auto, Res: 1K, 2K, auto

<!-- NOTES:grok-imagine-image-pro --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

seedream-4.5

Per-mode constraints:

• text_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 4K
• image_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 4K

Special Parameters:

• --sequential-image-generation <auto|disabled> — Control whether the model uses a more sequential generation path. Auto is the recommended balance unless you need to disable it for speed or compatibility. Default: auto.

Input Constraints:

• image: max 10MB, formats: jpg/jpeg/png/webp/bmp/tiff/gif, min 15px, max 36MP, aspect ratio range: 0.0625–16

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 4K
• image_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 4K
• route_c (0.75x)
• text_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 4K
• image_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 4K

<!-- NOTES:seedream-4.5 --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

seedream-5.0-lite

Per-mode constraints:

• text_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 3K
• image_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 3K

Special Parameters:

• --sequential-image-generation <auto|disabled> — Control whether the model uses a more sequential generation path. Auto is the recommended balance unless you need to disable it for speed or compatibility. Default: auto.
• --web-search — Let the model use live web results as extra context when the prompt depends on current facts, brands, places, or public references. Default: false.

Input Constraints:

• image: max 10MB, formats: jpg/jpeg/png/webp/bmp/tiff/gif, min 15px, max 36MP, aspect ratio range: 0.0625–16

Provider Routes:

• route_a (1x)
• text_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 3K
• image_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 3K
• route_c (0.75x)
• text_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 3K
• image_to_image: AR: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, Res: 2K, 3K

<!-- NOTES:seedream-5.0-lite --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

Video Models — Detailed Reference

Auto-generated from model-manifest.json. Hand-written notes in NOTES blocks are preserved.

veo-3.1

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, Res: 720p, 4k, 1080p, Duration: 4, 6, 8s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 4k, 1080p, Duration: 4, 6, 8s
• interpolation: AR: 16:9, 9:16, Res: 720p, 4k, 1080p, Duration: 8s
• reference_images: AR: 16:9, 9:16, Res: 720p, 4k, 1080p, Duration: 8s

Pricing:

• Base: 116 credits
• mode=text_to_video/image_to_video, duration=4, resolution=720p: 116 credits
• mode=text_to_video/image_to_video, duration=6, resolution=720p: 176 credits
• mode=text_to_video/image_to_video/interpolation/reference_images, duration=8, resolution=720p/1080p: 232 credits
• mode=text_to_video/image_to_video/interpolation/reference_images, duration=8, resolution=4k: 352 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, 4k, Dur: 4, 6, 8s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, 4k, Dur: 4, 6, 8s
• interpolation: AR: 16:9, 9:16, Res: 720p, 1080p, 4k, Dur: 8s
• reference_images: AR: 16:9, 9:16, Res: 720p, 1080p, 4k, Dur: 8s
• route_c (0.5x)
• text_to_video: AR: 16:9, 9:16, Res: 720p, 4k, Dur: 8s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 4k, Dur: 8s
• interpolation: AR: 16:9, 9:16, Res: 720p, 4k, Dur: 8s
• reference_images: AR: 16:9, 9:16, Res: 720p, 4k, Dur: 8s

Route-specific restrictions:

• route_c: when duration=4/6 — Route C only supports 8s
• route_c: when resolution=1080p — Route C does not support 1080p

<!-- NOTES:veo-3.1 --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

veo-3.1-fast

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, Res: 720p, 4k, 1080p, Duration: 4, 6, 8s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 4k, 1080p, Duration: 4, 6, 8s
• interpolation: AR: 16:9, 9:16, Res: 720p, 4k, 1080p, Duration: 8s

Pricing:

• Base: 44 credits
• mode=text_to_video/image_to_video, duration=4, resolution=720p: 44 credits
• mode=text_to_video/image_to_video, duration=6, resolution=720p: 68 credits
• mode=text_to_video/image_to_video/interpolation, duration=8, resolution=720p/1080p: 88 credits
• mode=text_to_video/image_to_video/interpolation, duration=8, resolution=4k: 192 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, 4k, Dur: 4, 6, 8s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, 4k, Dur: 4, 6, 8s
• interpolation: AR: 16:9, 9:16, Res: 720p, 1080p, 4k, Dur: 8s
• route_c (0.5x)
• text_to_video: AR: 16:9, 9:16, Res: 720p, 4k, Dur: 8s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 4k, Dur: 8s
• interpolation: AR: 16:9, 9:16, Res: 720p, 4k, Dur: 8s

Route-specific restrictions:

• route_c: when duration=4/6 — Route C only supports 8s
• route_c: when resolution=1080p — Route C does not support 1080p

<!-- NOTES:veo-3.1-fast --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

veo-3.1-lite

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, Duration: 4, 6, 8s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, Duration: 4, 6, 8s
• interpolation: AR: 16:9, 9:16, Res: 720p, 1080p, Duration: 4, 6, 8s

Pricing:

• Base: 22 credits
• duration=4, resolution=720p: 22 credits
• duration=6, resolution=720p: 33 credits
• duration=8, resolution=720p: 44 credits
• duration=8, resolution=1080p: 71 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, Dur: 4, 6, 8s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, Dur: 4, 6, 8s
• interpolation: AR: 16:9, 9:16, Res: 720p, 1080p, Dur: 4, 6, 8s

<!-- NOTES:veo-3.1-lite --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

sora-2

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, Res: 720p, Duration: 4, 8, 12, 16, 20s
• image_to_video: AR: 16:9, 9:16, Res: 720p, Duration: 4, 8, 12, 16, 20s

Pricing:

• Base: 40 credits
• duration=4: 40 credits
• duration=8: 80 credits
• duration=12: 120 credits
• duration=16: 160 credits
• duration=20: 200 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, Res: 720p, Dur: 4, 8, 12, 16, 20s
• image_to_video: AR: 16:9, 9:16, Res: 720p, Dur: 4, 8, 12, 16, 20s
• route_c (0.5x)
• text_to_video: AR: 16:9, 9:16, Res: 720p, Dur: 4, 8, 12s
• image_to_video: AR: 16:9, 9:16, Res: 720p, Dur: 4, 8, 12s

<!-- NOTES:sora-2 --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

sora-2-pro

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, 1024p, Duration: 4, 8, 12, 16, 20s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 1080p, 1024p, Duration: 4, 8, 12, 16, 20s

Pricing:

• Base: 120 credits
• 15 pricing tiers (varies by resolution/duration/mode)
• Range: 120–1400 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, Res: 720p, 1024p, 1080p, Dur: 4, 8, 12, 16, 20s
• image_to_video: AR: 16:9, 9:16, Res: 720p, 1024p, 1080p, Dur: 4, 8, 12, 16, 20s

<!-- NOTES:sora-2-pro --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

seedance-1.5-pro

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Duration: 4-12s
• image_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Duration: 4-12s
• interpolation: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Duration: 4-12s

Input Constraints:

• first_frame: max 30MB, formats: jpg/jpeg/png/webp/bmp/tiff/gif/heic/heif, min 300px, max 6000px, aspect ratio range: 0.4–2.5
• last_frame: max 30MB, formats: jpg/jpeg/png/webp/bmp/tiff/gif/heic/heif, min 300px, max 6000px, aspect ratio range: 0.4–2.5

Pricing:

• Base: 12 credits
• 27 pricing tiers (varies by resolution/duration/mode)
• Range: 12–144 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Dur: 4-12s
• image_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Dur: 4-12s
• interpolation: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Dur: 4-12s
• route_c (1x)
• text_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Dur: 4-12s
• image_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Dur: 4-12s
• interpolation: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, 1080p, Dur: 4-12s

<!-- NOTES:seedance-1.5-pro --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

seedance-2.0

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• image_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• interpolation: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• reference_images: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• multimodal_reference: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• video_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• video_extension: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s

Special Parameters:

• --audio — Generate audio together with the video when supported by the model. Default: true.
• --web-search — Allow the model to use web search when the provider supports it. Default: false.

Input Constraints:

• first_frame: max 30MB, formats: jpg/jpeg/png/webp/bmp/tiff/gif, min 300px, max 6000px, aspect ratio range: 0.4–2.5
• last_frame: max 30MB, formats: jpg/jpeg/png/webp/bmp/tiff/gif, min 300px, max 6000px, aspect ratio range: 0.4–2.5
• video: max 50MB, formats: mp4/mov, min 300px, max 6000px, min 2s, max 15s
• audio: max 15MB, formats: wav/mp3, min 2s, max 15s

Pricing:

• Base: 37 credits
• 24 pricing tiers (varies by resolution/duration/mode)
• Range: 30–238 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• image_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• interpolation: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• reference_images: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• multimodal_reference: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• video_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• video_extension: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s

Validation Rules:

• When mode=text_to_video/image_to_video/interpolation/reference_images/video_to_video/video_extension: forbid reference_videos — reference_videos are only available in multimodal_reference mode
• When mode=text_to_video/image_to_video/interpolation/reference_images/video_to_video/video_extension: forbid reference_audios — reference_audios are only available in multimodal_reference mode
• When mode=multimodal_reference: forbid first_frame — multimodal_reference uses reference_images instead of first_frame
• When mode=multimodal_reference: forbid last_frame — multimodal_reference does not support last_frame
• When mode=multimodal_reference: forbid video — multimodal_reference uses reference_videos instead of video

<!-- NOTES:seedance-2.0 --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

seedance-2.0-fast

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• image_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• interpolation: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• reference_images: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• multimodal_reference: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• video_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s
• video_extension: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Duration: 4-15s

Special Parameters:

• --audio — Generate audio together with the video when supported by the model. Default: true.
• --web-search — Allow the model to use web search when the provider supports it. Default: false.

Input Constraints:

• first_frame: max 30MB, formats: jpg/jpeg/png/webp/bmp/tiff/gif, min 300px, max 6000px, aspect ratio range: 0.4–2.5
• last_frame: max 30MB, formats: jpg/jpeg/png/webp/bmp/tiff/gif, min 300px, max 6000px, aspect ratio range: 0.4–2.5
• video: max 50MB, formats: mp4/mov, min 300px, max 6000px, min 2s, max 15s
• audio: max 15MB, formats: wav/mp3, min 2s, max 15s

Pricing:

• Base: 30 credits
• 24 pricing tiers (varies by resolution/duration/mode)
• Range: 25–192 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• image_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• interpolation: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• reference_images: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• multimodal_reference: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• video_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s
• video_extension: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive, Res: 480p, 720p, Dur: 4-15s

Validation Rules:

• When mode=text_to_video/image_to_video/interpolation/reference_images/video_to_video/video_extension: forbid reference_videos — reference_videos are only available in multimodal_reference mode
• When mode=text_to_video/image_to_video/interpolation/reference_images/video_to_video/video_extension: forbid reference_audios — reference_audios are only available in multimodal_reference mode
• When mode=multimodal_reference: forbid first_frame — multimodal_reference uses reference_images instead of first_frame
• When mode=multimodal_reference: forbid last_frame — multimodal_reference does not support last_frame
• When mode=multimodal_reference: forbid video — multimodal_reference uses reference_videos instead of video

<!-- NOTES:seedance-2.0-fast --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

grok-imagine-video

Per-mode constraints:

• text_to_video: AR: 2:3, 3:2, 1:1, 16:9, 9:16, 4:3, 3:4, Res: 720p, 480p, Duration: 1-15s
• image_to_video: AR: 2:3, 3:2, 1:1, 16:9, 9:16, 4:3, 3:4, auto, Res: 720p, 480p, Duration: 1-15s
• reference_images: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, Res: 480p, 720p, Duration: 1-10s
• video_to_video: AR: 16:9, Res: 720p, Duration: 5s
• video_extension: AR: 16:9, Res: 720p, Duration: 2-10s

Per-mode limits:

• video_to_video: derived (auto-set): aspect_ratio, resolution, duration
• video_extension: derived (auto-set): aspect_ratio, resolution

Input Constraints:

• video: formats: mp4, max 8.7s

Pricing:

• Base: 1 credits
• 30 pricing tiers (varies by resolution/duration/mode)
• Range: 6–120 credits

Provider Routes:

• route_a (1x)
• text_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, Res: 480p, 720p, Dur: 1-15s
• image_to_video: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, auto, Res: 480p, 720p, Dur: 1-15s
• reference_images: AR: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, Res: 480p, 720p, Dur: 1-10s
• video_to_video: AR: 16:9, Res: 720p, Dur: 5s
• video_extension: AR: 16:9, Res: 720p, Dur: 2-10s
• route_c (0.5x)
• text_to_video: AR: 2:3, 3:2, 1:1, Res: 720p, Dur: 10s
• image_to_video: AR: 2:3, 3:2, 1:1, Res: 720p, Dur: 10s

Route-specific restrictions:

• route_c: when mode=reference_images/video_to_video/video_extension — Route C is not available for this mode
• route_c: when aspect_ratio=16:9/9:16/4:3/3:4/auto — Route C only supports 1:1, 3:2, and 2:3 aspect ratios
• route_c: when resolution=480p — Route C only supports 720p
• route_c: when duration=1/2/3/4/5/6/7/8/9/11/12/13/14/15 — Route C only supports 10s

Validation Rules:

• When mode=video_to_video: forbid aspect_ratio — video_to_video follows the source video aspect ratio
• When mode=video_to_video: forbid resolution — video_to_video follows the source video resolution
• When mode=video_extension: forbid aspect_ratio — video_extension follows the source video aspect ratio
• When mode=video_extension: forbid resolution — video_extension follows the source video resolution

<!-- NOTES:grok-imagine-video --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

kling-2.6

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, 1:1, Res: 720p, 1080p, Duration: 5, 10s
• image_to_video: AR: 16:9, Res: 720p, 1080p, Duration: 5, 10s
• motion_control: AR: 16:9, Res: 720p, 1080p, Duration: 5s
• interpolation: AR: 16:9, Res: 1080p, Duration: 5, 10s

Special Parameters:

• --audio — Generate audio together with the video when supported by the model. Default: false.
• --character-orientation <image|video> — Choose whether character motion and framing should follow the uploaded image or the source video. Required.
• --keep-original-audio — Keep the original soundtrack from the source video instead of replacing or dropping it. Default: true.

Input Constraints:

• prompt: max 2500 chars
• first_frame: max 10MB, formats: jpg/jpeg/png, min 300px, max 65536px, aspect ratio range: 0.4–2.5
• last_frame: max 10MB, formats: jpg/jpeg/png, min 300px, max 65536px, aspect ratio range: 0.4–2.5
• video: max 100MB, formats: mp4/mov, min 340px, max 3850px, min 3s, max 30s, when character_orientation=video: max 30s, when character_orientation=image: max 10s

Pricing:

• Base: 26 credits
• mode=image_to_video/text_to_video, resolution=720p, duration=5: 26 credits
• mode=image_to_video/text_to_video, resolution=720p, duration=10: 51 credits
• mode=image_to_video/interpolation/text_to_video, resolution=1080p, duration=5: 40 credits
• mode=image_to_video/interpolation/text_to_video, resolution=1080p, duration=10: 80 credits
• Surcharge: audio=true adds +40 credits
• Surcharge: audio=true adds +80 credits

Provider Routes:

• route_c (0.5x)
• text_to_video: AR: 16:9, 9:16, 1:1, Res: 720p, 1080p, Dur: 5, 10s
• image_to_video: AR: 16:9, Res: 720p, 1080p, Dur: 5, 10s
• motion_control: AR: 16:9, Res: 720p, 1080p, Dur: 5s
• interpolation: AR: 16:9, Res: 1080p, Dur: 5, 10s

Validation Rules:

• When mode=text_to_video/image_to_video/interpolation, resolution=720p: forbid audio — 720p does not support audio
• When mode=image_to_video: forbid aspect_ratio — aspect_ratio follows first_frame in image_to_video
• When mode=image_to_video: forbid last_frame — use mode=interpolation for last_frame
• When has last_frame: require resolution=1080p — last_frame requires resolution=1080p
• When has last_frame: forbid audio — last_frame does not support audio
• When mode=motion_control: forbid last_frame
• When mode=motion_control: forbid duration
• When mode=motion_control: forbid aspect_ratio
• When mode=motion_control: force audio=null — audio is not applicable to motion_control

<!-- NOTES:kling-2.6 --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

kling-3.0

Per-mode constraints:

• text_to_video: AR: 16:9, 9:16, 1:1, Res: 720p, 1080p, Duration: 3-15s
• image_to_video: AR: 16:9, Res: 720p, 1080p, Duration: 3-15s
• interpolation: AR: 16:9, Res: 720p, 1080p, Duration: 3-15s
• motion_control: AR: 16:9, Res: 720p, 1080p, Duration: 3s

Special Parameters:

• --audio — Generate audio together with the video when supported by the model. Default: false.
• --multi-shot — Split one generation into multiple consecutive shots so you can build a sequence instead of a single continuous clip. Default: false.
• --shot-type <customize|intelligence> — Choose how multi-shot scenes are controlled. Customize uses per-shot instructions, while intelligence lets the model plan the sequence from one prompt.
• --multi-prompt <json> — Provide the prompt and duration for each individual shot when multi-shot customize mode is enabled.

Input Constraints:

• prompt: max 2500 chars
• first_frame: max 10MB, formats: jpg/jpeg/png, min 300px, max 65536px, aspect ratio range: 0.4–2.5
• last_frame: max 10MB, formats: jpg/jpeg/png, min 300px, max 65536px, aspect ratio range: 0.4–2.5

Pricing:

• Base: 30 credits
• 26 pricing tiers (varies by resolution/duration/mode)
• Range: 30–185 credits
• Surcharge: audio=true adds +14 credits
• Surcharge: audio=true adds +18 credits
• Surcharge: audio=true adds +23 credits
• Surcharge: audio=true adds +27 credits
• Surcharge: audio=true adds +32 credits
• Surcharge: audio=true adds +36 credits
• Surcharge: audio=true adds +41 credits
• Surcharge: audio=true adds +45 credits
• Surcharge: audio=true adds +50 credits
• Surcharge: audio=true adds +54 credits
• Surcharge: audio=true adds +59 credits
• Surcharge: audio=true adds +63 credits
• Surcharge: audio=true adds +68 credits
• Surcharge: audio=true adds +11 credits
• Surcharge: audio=true adds +14 credits
• Surcharge: audio=true adds +18 credits
• Surcharge: audio=true adds +22 credits
• Surcharge: audio=true adds +25 credits
• Surcharge: audio=true adds +29 credits
• Surcharge: audio=true adds +33 credits
• Surcharge: audio=true adds +37 credits
• Surcharge: audio=true adds +40 credits
• Surcharge: audio=true adds +44 credits
• Surcharge: audio=true adds +48 credits
• Surcharge: audio=true adds +51 credits
• Surcharge: audio=true adds +55 credits

Provider Routes:

• route_c (1x)
• text_to_video: AR: 16:9, 9:16, 1:1, Res: 720p, 1080p, Dur: 3-15s
• image_to_video: AR: 16:9, Res: 720p, 1080p, Dur: 3-15s
• interpolation: AR: 16:9, Res: 720p, 1080p, Dur: 3-15s
• motion_control: AR: 16:9, Res: 720p, 1080p, Dur: 3s

Validation Rules:

• When mode=image_to_video, has aspect_ratio: forbid aspect_ratio — aspect ratio follows first_frame
• When mode=image_to_video, has last_frame: forbid last_frame — use mode=interpolation for last_frame
• When has shot_type, multi_shot=false: forbid shot_type — shot_type only valid when multi_shot=true
• When has multi_prompt, multi_shot=false: forbid multi_prompt — multi_prompt only valid when multi_shot=true
• When multi_shot=false: require prompt — prompt required when multi_shot=false
• When multi_shot=true: require shot_type — shot_type required when multi_shot=true
• When has prompt, multi_shot=true, shot_type=customize: forbid prompt — use multi_prompt instead of prompt
• When multi_shot=true, shot_type=customize: require multi_prompt — multi_prompt required for customize mode
• When has multi_prompt, multi_shot=true, shot_type=intelligence: forbid multi_prompt — intelligence mode does not allow multi_prompt
• When multi_shot=true, shot_type=intelligence: require prompt — prompt required for intelligence mode

<!-- NOTES:kling-3.0 --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

kling-2.6-motion-control

Per-mode constraints:

• motion_control: AR: 16:9, Res: 720p, 1080p, Duration: 5s

Special Parameters:

• --audio — Generate audio together with the video when supported by the model. Default: false.
• --character-orientation <image|video> — Choose whether character motion and framing should follow the uploaded image or the source video. Required.
• --keep-original-audio — Keep the original soundtrack from the source video instead of replacing or dropping it. Default: true.

Input Constraints:

• prompt: max 2500 chars
• first_frame: max 10MB, formats: jpg/jpeg/png, min 300px, max 65536px, aspect ratio range: 0.4–2.5
• video: max 100MB, formats: mp4/mov, min 340px, max 3850px, min 3s, max 30s, when character_orientation=video: max 30s, when character_orientation=image: max 10s

Pricing:

• Base: 8 credits
• Unit: per_second
• resolution=720p: 8 credits
• resolution=1080p: 12 credits

Provider Routes:

• route_c (0.5x)
• motion_control: AR: 16:9, Res: 720p, 1080p, Dur: 5s

<!-- NOTES:kling-2.6-motion-control --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---

kling-3.0-motion-control

Per-mode constraints:

• motion_control: AR: 16:9, Res: 720p, 1080p, Duration: 3s

Special Parameters:

• --audio — Generate audio together with the video when supported by the model. Default: false.
• --character-orientation <image|video> — Choose whether character motion and framing should follow the uploaded image or the source video. Required.
• --keep-original-audio — Keep the original soundtrack from the source video instead of replacing or dropping it. Default: true.

Input Constraints:

• prompt: max 2500 chars
• first_frame: max 10MB, formats: jpg/jpeg/png, min 300px, max 65536px, aspect ratio range: 0.4–2.5
• video: max 100MB, formats: mp4/mov, min 340px, max 3850px, min 3s, max 30s, when character_orientation=video: max 30s, when character_orientation=image: max 10s

Pricing:

• Base: 14.5 credits
• Unit: per_second
• resolution=720p: 14.5 credits
• resolution=1080p: 16 credits

Provider Routes:

• route_c (1x)
• motion_control: AR: 16:9, Res: 720p, 1080p, Dur: 3s

Validation Rules:

• When mode=motion_control: forbid last_frame
• When mode=motion_control: forbid duration
• When mode=motion_control: forbid aspect_ratio
• When mode=motion_control: force audio=null — audio is not applicable to motion_control

<!-- NOTES:kling-3.0-motion-control --> <!-- Add hand-written tips and best practices here --> <!-- NOTES:END -->

---