Try Seedance 2.5 image-to-video with first-frame and first/last-frame control. Compare real 720p results, copy tested prompts, review 4–30 second pricing, and call POST /v1/tasks.
Provider: ByteDance
Category: video generation
Endpoint: /v1/tasks
Status: Available
Cost: See live page pricing
Seedance 2.5 image-to-video creates up to 30 seconds of video from a first frame or a first/last frame pair, with the output ratio following the first frame.
Pricing
Standard Usage
Use images to control the first and last frame.
jpeg, png, webp, bmp, tiff, or gif; up to 30MB; 300-6000 px per side and a 0.4-2.5 aspect ratio
jpeg, png, webp, bmp, tiff, or gif; up to 30MB; 300-6000 px per side and a 0.4-2.5 aspect ratio
Choose 4-30 seconds in 1-second steps
Up to 10 clips; WAV or MP3; up to 15MB and 2-30 seconds per clip; combined duration up to 30 seconds; Up to 10 items
Choose the video format. MP4 is broadly compatible; MOV is intended for professional post-production
Generate synchronized audio for the video
Whether to add an AI-generated watermark
Use web search signals for generation
Ready for generation
Configure your parameters and click "Run" to see the output here.
Real API image-to-video results
Four real runs cover a first-frame camera move, first/last-frame lighting transition, growth sequence, and square character animation. Inputs and outputs remain separate and complete, with copyable prompts for reproduction or adaptation.
A single source still drives a restrained dolly-in while steam, morning light, and shallow depth of field remain coherent.
Prompt used for this verified run
Slow, steady dolly-in on the steaming coffee cup, camera glides forward and slightly down, steam swirls and drifts upward in soft morning light, gentle handheld micro-motion, cinematic shallow depth of field, warm natural color grade
The endpoints lock composition while daylight fades, the lamp and laptop wake up, and city lights appear without camera movement.
Prompt used for this verified run
Time-lapse from bright daytime to night at the same desk: daylight fades to warm lamp glow, laptop screen turns on and begins glowing softly, city lights flicker on outside the window, smooth continuous lighting transition, no camera movement, static locked-off shot
A locked camera and stable morning light let the model concentrate on a readable sprouting and leaf-unfurling sequence.
Prompt used for this verified run
Time-lapse growth of a basil seedling sprouting and unfurling leaves inside the terracotta pot, smooth continuous motion, soft morning light staying consistent, camera locked off, no cuts
A square green-screen source tests silhouette stability, alternating leg motion, and a repeatable 2D character cycle.
Prompt used for this verified run
side view of a cute cartoon Samoyed dog character walking cycle animation, smooth looping walk motion, character walking in place from left to right, seamless loop with first and last frames perfectly matching, bright solid green screen background for chroma key, 2D animation style, clean cel shading, professional game sprite animation, high quality, consistent character design, full body visible
Choose by starting material, reference method, resolution, and shot-control requirements.
Start from text
Use the text route when no source frame is required and the composition should be created from the prompt.
Guide with reference video
Use one to ten reference videos when motion, performance, or camera language must be transferred.
Higher-resolution alternative
A one-image alternative for 720p, 1080p, and 4K output with native audio.
Seedance 2.5 Image to Video can extend one first frame into a shot, use first and last frames to lock both endpoints, or switch to reference-image and reference-audio guidance. It supports four to thirty seconds, 480p/720p, adaptive framing, and synchronized audio for product motion, lighting transitions, growth sequences, and character animation.
The Playground and API use the same model ID: seedance-2.5/image-to-video. Endpoint images define where the shot starts and ends, while the prompt directs motion, camera, lighting, and sound between them.
These specifications reflect the current HiAPI Seedance 2.5 Image to Video request contract. Endpoint frames and multimodal references are separate input modes and should not be mixed.
480p is 279 Credits / second; 720p is about 603 Credits / second.
The shared three-step shell is adapted for first-frame and endpoint control, keeping different reference modes out of the same request.
This is a minimal asynchronous first/last-frame request. Use the API docs for callbacks, task retrieval, reference limits, and error handling.
curl --request POST 'https://api.hiapi.ai/v1/tasks' \
--header 'Authorization: Bearer YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"model": "seedance-2.5/image-to-video",
"input": {
"prompt": "Time-lapse growth of a basil seedling, smooth continuous motion, soft morning light, locked camera, no cuts.",
"first_frame_url": "https://static.hiapi.ai/blog/seedance-2-5-image-to-video-prompt-recipes/first-frame.jpg",
"last_frame_url": "https://static.hiapi.ai/blog/seedance-2-5-image-to-video-prompt-recipes/last-frame.jpg",
"resolution": "720p",
"duration": 4,
"aspect_ratio": "adaptive",
"generate_audio": true
}
}'No. In endpoint-frame mode, first_frame_url alone extends the source into a complete shot; add last_frame_url when the ending state must be controlled. A last frame cannot be used by itself. You can instead use reference_image_urls or reference_audio_urls as a separate reference mode without endpoint frames.
The current HiAPI Playground treats them as separate input modes: use first/last frames, or use reference_image_urls and reference_audio_urls. Keep one mode to avoid conflicting guidance. Use seedance-2.5/reference-to-video when a reference video must guide motion or camera movement.
The current request contract accepts integer durations from 4 to 30 seconds at 480p or 720p. aspect_ratio is adaptive and follows the first frame. The four real cases on this page verify 720p at four and five seconds; 30 seconds is a supported contract value, not mislabeled as a tested page example.
Billing scales with output duration and resolution, with separate per-second tiers for 480p and 720p. Check the live model price and task record for the final charge.
Describe the continuous change between the endpoints, then lock camera behavior, lighting, and details that must remain stable. A day-to-night shot can specify time passing and lights switching on; a growth shot can define the growth order, locked camera, and no cuts. The images define the endpoints while the prompt directs the path between them.
Yes. generate_audio defaults to true and can create synchronized dialogue, ambience, sound effects, and background audio. Set it to false for a silent result. Review volume, dialogue accuracy, and music rights before publishing commercial work.
Send an asynchronous task to POST /v1/tasks, set model to seedance-2.5/image-to-video, and pass prompt, first_frame_url, an optional last_frame_url, resolution, duration, and generate_audio in input. The create request returns a taskId. See the API docs for callbacks, retrieval, and complete media constraints.
The first frame defines subject, composition, and shape. Add a matching last frame only when the ending state must be fixed, leaving room for the intended motion.
Write the middle process in time order, state whether the camera is locked, preserve stable details, and place sound cues next to their matching events.
Start at four seconds to inspect endpoint continuity and subject stability, then increase duration or create the production task with seedance-2.5/image-to-video.