SD Video — video with sound from one prompt, built on MiniMax H3
SD Video is SeedRouter's video model, built on MiniMax H3: describe a scene and get a 5–15 second clip with its own dialogue, ambience and effects, from $0.0200 per second.
- Text to video
- Image to video
- Reference to video
- Generated audio
- Repeatable seeds
- Six aspect ratios
Try SD Video before you write any code
Run SD Video with the exact request your code will send. Every parameter the API takes is here.
sd-videofrom $0.0200/s
Pick a mode, describe the shot, add a first frame or references if the mode needs them, and submit once. The clip appears when the task finishes.
The shot and its sound: subject, place, camera, light, dialogue and effects. Up to 32,000 characters.
Start from text alone, from a first frame, or from your own images, videos and audio. References is the default.
Files are uploaded and sent as URLs.
Up to 9. Call them <Picture 1>, <Picture 2> in the prompt.
Files are uploaded and sent as URLs.
Up to 3, called <Video 1> and on. Each adds up to 5 billed seconds.
Files are uploaded and sent as URLs.
Up to 3, called <Audio 1> and on. 12 references at most in total.
768p is the default; 480p costs less.
Default is 16:9 for text and adaptive for references, which follows your first reference. A first frame sets its own shape.
5–15 seconds. Default is 5.
turbo (default) expands your prompt fast, quality takes longer, disabled sends it word for word.
Optional, 0–4294967295. Keep it fixed to repeat a shot while you change one detail.
- reference_images: Provide at least one reference image or video.
Estimated at current rates from the mode, resolution and length. In reference mode, reference video seconds are added once the files load.
Try one of these prompts
Sample clips with their generated audio
SD Video pricing
SD Video is billed per second of video. The rate depends on the mode and the resolution, and every rate below is the one that bills your account.
- SD Video
- $0.15/clip
- SD Video
- 666seconds
All SD Video prices come from the rates that bill your account. Failed requests are not charged.
| Mode | Billed seconds | Rate | Billing |
|---|---|---|---|
| Text to video · 480p | Seconds of the delivered video | $0.0200 / s | Charged from the finished video |
| Text to video · 768p | Seconds of the delivered video | $0.0300 / s | Charged from the finished video |
| Image to video · 480p | Seconds of the delivered video | $0.0200 / s | Charged from the finished video |
| Image to video · 768p | Seconds of the delivered video | $0.0300 / s | Charged from the finished video |
| Reference to video · 480p | Output seconds plus up to 5 seconds per reference video; images and audio are free | $0.0400 / s | Charged from the finished video |
| Reference to video · 768p | Output seconds plus up to 5 seconds per reference video; images and audio are free | $0.0600 / s | Charged from the finished video |
| Failed requests | Any task that ends failed | Free | Nothing is charged |
What is SD Video?
SD Video is SeedRouter's video generation model, built on MiniMax H3. One SD Video call makes the whole scene at once: subject, setting, light and sound, including spoken lines, room tone and effects that match the space you describe.
You can call SD Video three ways: from a text prompt, from an image that becomes the first frame, or from up to 12 reference images, videos and audio clips that the shot must keep. Submit once, get a task ID, and collect the MP4 when it is ready.
SD Video output is MP4 (H.264) at 24 fps with 32 kHz stereo AAC audio. Clips run 5–15 seconds at 480p or 768p, for example 1344 × 768 at 16:9.
Two ways to use SD Video
Call the API from your backend, or hand the job to a coding agent.
Call the API
Send the SD Video request from your server, keep the task ID, and read the finished clip from the task.
- 1Create an API key
- 2POST the request with your prompt
- 3Poll the task until the video is ready
Let an agent do it
Copy a ready prompt into Claude Code, Codex or Cursor. It asks before spending and downloads the result.
- 1Copy the prompt
- 2Paste it into your agent and describe the shot
- 3Approve the cost, then get the video
What SD Video is good at
SD Video is at its best on short, contained shots where picture and sound have to agree.
Sound in the same render
Dialogue, ambience and effects come with every SD Video clip. No voiceover pass, no syncing.
Scenes that hold together
SD Video keeps objects at their size and place; nothing duplicates or vanishes mid-shot.
Does what you asked
SD Video follows a brief literally. No dialogue in the prompt means no dialogue in the clip.
Keeps your product
Give SD Video a photo as a reference and its colour, shape and markings carry into the shot.
Short titles in shot
Spell a short phrase out exactly in the prompt and SD Video renders it on a wall, a sign or a page.
Repeatable shots
Fix the seed and SD Video returns the same clip for the same prompt, so you can change one detail at a time.
What teams use SD Video for
Talking scenes
A person speaks a line to camera, with the room sound to match, from one SD Video request.
Product shots
Put your real product into a new scene with SD Video reference mode instead of a photo shoot.
Social clips
SD Video clips in vertical 9:16 for Reels and TikTok or wide 16:9 for YouTube, sound already in.
Training videos
Short SD Video explainer shots where the speech carries the steps and the camera stays still.
Title cards
Generate an SD Video shot with your phrase painted, stamped or printed inside it.
Animated styles
Name a real medium, such as screen print or clay, and SD Video renders the shot in it.
Start with SD Video in four steps
- 01
Create a key
Sign in, top up any amount, and create an API key. Credits never expire.
- 02
Send one request
POST the SD Video body to /v1/videos/generations with your prompt and any media URLs.
- 03
Poll the task
Keep the SD Video task ID and check it every 10–20 seconds until the status is completed.
- 04
Download the video
Save the hosted SD Video URL from the task to your own storage.
SD Video at a glance
The facts that decide whether SD Video fits the job.
| Feature | SD Video |
|---|---|
| Resolution | 480p, 768p |
| Clip length | 5–15 seconds |
| Inputs | Text, a first frame, or images, videos and audio |
| Audio | Always generated: dialogue, ambience, effects |
| Repeatable | Yes, with a fixed seed |
| Billing | Per second, by mode and resolution |
Need 1080p, or clips longer than 15 seconds? See Seedance 2.5 and Seedance 2.0.
What to know before you start
Best on contained shots
One subject, one place, one action and a still camera work best. Long on-screen text, soft motion like hair or paper, and close hand work are the weak spots.
Media is sent as URLs
Send SD Video images, clips and audio as public HTTP(S) links, not base64. The Playground uploads for you.
Results arrive by task
An SD Video request returns a task ID first. Poll it until the video is ready.
Tasks cannot be cancelled
Once submitted, an SD Video task runs to the end. Check the parameters before you send it.
Why run SD Video on SeedRouter
Pay per second of video
SD Video is billed on the length of the finished clip, at the rates in the table above. No subscription, no minimum.
Failed requests are free
If an SD Video task ends failed, nothing is charged. You only pay for videos that are delivered.
Credits never expire
Top up once and use the balance whenever you need it.
Hosted results
Finished SD Video clips come back as URLs on our CDN. Keep a copy of anything you need long term.
One balance for every model
The same top-up pays for video, image and text models in the catalogue.
Other models on SeedRouter
Same balance, same task-based workflow.
SD Video guides
Practical guides to choosing, pricing and integrating the model.
MiniMax H3 costs $0.08 a second at 768p and $0.13 at 2K on the official API. Its open weights are free, need about 108 GB and exclude the US, EU, UK and Korea.
How to write MiniMax H3 prompts: the official three-part structure, dialogue tags, shot timing and reference labels, tested on four clips.
MiniMax H3 vs Seedance 2.5: a blind ranking calls them tied. Seedance 2.5 leads on anatomy, lip sync and 30-second clips; H3 costs far less per minute.
SD Video FAQ
What is SD Video?+
SD Video is SeedRouter's video generation model, built on MiniMax H3. It makes one clip of 5 to 15 seconds at 480p or 768p, with dialogue, ambience and effects generated together with the picture. You can start from a text prompt, a first-frame image, or up to 12 reference images, videos and audio clips.
Is SD Video free?+
No. SD Video is paid per second of finished video, and new accounts get a small credit to try it. There is no subscription: you top up any amount, credits never expire, and failed requests are not charged.
How much does SD Video cost?+
A 5-second 768p text-to-video clip costs $0.15. SD Video is billed per second at a rate set by the mode and resolution, and reference mode also bills each reference video, up to 5 seconds per video. The pricing table on this page shows every current rate.
Does SD Video need a monthly subscription?+
No. SD Video is pay as you go. Add credit once, spend it on any model in the catalogue, and top up again when you need more.
Does SD Video generate audio?+
Yes. Every SD Video clip comes with a stereo audio track generated from the same prompt: spoken lines, room tone and effects. Write the dialogue into the prompt, about 2.5 words per second, and say no music if you do not want a music bed.
How do references work in SD Video?+
In reference mode, SD Video takes up to 9 images, 3 videos and 3 audio clips, 12 in total. Refer to them in the prompt by position, such as <Picture 1> or <Video 1>, and describe the scene around them. To start a clip on an exact image, use image-to-video mode instead.
How do I get an API key for SD Video?+
Create a SeedRouter account, add credit, and create a key in the dashboard. Use it as a Bearer token on the SD Video request to /v1/videos/generations; the same key works for every model.
How do I write a good SD Video prompt?+
Write long prompts: a few hundred characters or more gives SD Video much better output. Name the camera and lens instead of a mood, describe the sound, spell out any words that should appear, and rule out logos you did not ask for.
Generate your first SD Video clip
Try SD Video in the Playground, then move the same request into your code.