← Back to pipelines

Video Generator

Generate AI videos from a text prompt, image, or reference clip. Cinematic shots, product reveals, lifestyle ads, with synchronized audio and natural motion.

9.5

Still needed: Prompt

Runs on

API

Examples

Reference → ResultSeedance 2 Pro

Pyra, cinematic (Seedance 2)

Seedance 2 Pro. Same Pyra, ultrawide: a native 21:9 cinematic wide-frame glide with sweeping camera motion, the richer result Veo's vertical draft can't produce.

Reproduce

Best for

  • •Creating short video clips from text descriptions or images
  • •Product reveals and scene transitions with real camera movement
  • •Cinematic shots that need real motion (not just pan/zoom on a still)
  • •Text-to-video with native audio on models that support it (others are silent, add narration in video-reel)

Tips

  • ✓Provide detailed text prompts describing the scene, motion, and desired audio
  • ✓Use a start image for consistent framing, the model animates from it
  • ✓Works best for product reveals, scene changes, and dramatic transformations
  • ✓For simple pan/zoom on stills, image-motion is 10x faster and cheaper
  • ✓Native audio is generated alongside the video on models that support it, no separate voiceover step. Other models are silent; add narration or music in video-reel.

Recipes using this pipeline

Articles about this pipeline

Video Generator API

Call this pipeline from your own code. One request dispatches a run; the model is an input field, not a separate endpoint.

11 inputs
1 required
from 9.5 credits
bun add @pipe2-ai/sdk 
import { createClient } from "@pipe2-ai/sdk";

const client = createClient(process.env.PIPE2_TOKEN!);

const { run_pipeline } = await client.RunPipeline({
  pipeline_slug: "video-generator",
  input: {
    prompt: "A detective dials a rotary phone in a dimly lit alleyway, film noir style",
    model: "veo-3-1-lite",
  },
});
pip install pipe2 
import asyncio
import os

from pipe2 import Pipe2Client

client = Pipe2Client(token=os.environ["PIPE2_TOKEN"])

run = asyncio.run(client.run_pipeline(
    pipeline_slug="video-generator",
    input={
        "prompt": "A detective dials a rotary phone in a dimly lit alleyway, film noir style",
        "model": "veo-3-1-lite",
    },
))
go get github.com/pipe2-ai/sdk-go 
client := pipe2.NewClient(os.Getenv("PIPE2_TOKEN"))

input := json.RawMessage(`{
  "prompt": "A detective dials a rotary phone in a dimly lit alleyway, film noir style",
  "model": "veo-3-1-lite"
}`)

run, err := pipe2.RunPipeline(context.Background(), client, "video-generator", input)
if err != nil {
    log.Fatal(err)
}
brew install pipe2-ai/tap/pipe2 
pipe2 pipelines run \
  --pipeline video-generator \
  --input-json '{"prompt": "A detective dials a rotary phone in a dimly lit alleyway, film noir style", "model": "veo-3-1-lite"}' \
  --wait

Get an API token →

Parameters

prompt required

Text description of the video. The AI enhances this automatically with cinematic details.

string
aspect_ratio

Auto lets the director infer orientation from your prompt. Veo supports 16:9 and 9:16; 1:1, 4:3, 3:4 and 21:9 run on Seedance.

auto16:99:161:14:33:421:9 default auto
audio

Generate a synchronised audio track with the video.

boolean default true
Show all 11 inputs
duration

Auto lets the director right-size length from your prompt. Veo supports 4, 6 and 8s; 10s and longer run on Seedance.

auto468101215 default auto
end_image

Optional: URL of an image to use as the last frame. AI interpolates between start and end.

string
model

auto: a fast, flat-priced Veo 3.1 Lite draft; escalates to Seedance Pro for references, 10s+, or square/classic ratios. Pin a Veo tier (Lite / Fast / Standard) or a Seedance tier (Pro / Fast / Mini) to override.

autoveo-3-1-liteveo-3-1-fastveo-3-1-standardseedance-2-0-miniseedance-2-0-fastseedance-2-0-pro default auto
reference_audio_url

Optional: audio to use as the background track for the generated video. Auto-routes to Seedance Pro.

string
reference_images

Optional: up to 4 reference images. With 1-2 images Veo can use them as start and end frames; more images route automatically to Seedance Pro.

array
reference_video_url

Optional: a reference video for camera movement, pacing, or composition. Auto-routes the run to Seedance Pro.

string
resolution

Engine-aware: Veo has no 480p (Fast/Standard add 4K); Seedance Mini/Fast cap at 720p, Pro adds 1080p/4K.

480p720p1080p4k default 720p
seed

Optional: pin the random seed for a reproducible Seedance result. Leave empty for a fresh random seed each run.

integer

Use it from an AI agent

Point any MCP client at pipe2 and your agent gets these tools. It finds this pipeline, reads the same schema above, then runs it.

  • list_pipelines
  • get_pipeline_schema
  • run_pipeline
  • get_pipeline_run_status
  • request_upload
https://mcp.pipe2.ai/mcp 
{
  "mcpServers": {
    "pipe2ai": {
      "url": "https://mcp.pipe2.ai/mcp",
      "headers": { "Authorization": "Bearer YOUR_TOKEN" }
    }
  }
}

Full API reference MCP setup CLI reference

AI Video Generator

Generate short videos from a text prompt, image, or reference clip. Cinematic narrative shots, product ads, social-ready vertical content, with synchronized audio, real camera movement, and natural motion.

What you can make

  • Text-to-video scenes: describe a moment and get a finished clip with audio
  • Image-to-video: start from a reference image and animate it into a cinematic shot
  • Product and lifestyle ads: bring a reference video for camera move and a reference audio for the music bed
  • Vertical shorts: 9:16 for TikTok, Reels, and Shorts in one call
  • Cinematic explainers: golden hour, film noir, anime, watercolor, 35mm film, any style on demand

How it works

  1. Describe the scene: subject, action, camera, lighting, style
  2. Optionally attach references: images, a reference video for camera movement, or audio for the soundtrack
  3. Pick aspect and length: 16:9, 9:16, 1:1, 4:3, 3:4, or 21:9, 4–15 seconds
  4. Generate: the right AI engine is picked automatically based on what you uploaded

Prompt tips

  • Include camera angles: close-up, aerial, tracking shot, dolly in
  • Describe lighting: golden hour, neon glow, film noir shadows, volumetric fog
  • Set the style: cinematic, anime, watercolor, 35mm film look
  • Mention audio in a separate sentence, "rain pattering on concrete, distant thunder"
  • For longer clips with reference video and audio, the multimodal path keeps your soundtrack and camera move in sync

Frequently Asked Questions

How long are generated videos?
4 to 15 seconds per clip. Veo caps at 8s, so 10s and longer route automatically to Seedance. Combine multiple clips into longer sequences with the Video Reel pipeline.
Does it generate audio?
Yes. Veo synthesizes audio in sync with the video, describe the sounds you want in your prompt. Seedance can add a soundtrack too, or switch the audio toggle off for a silent clip.
What aspect ratios are supported?
16:9, 9:16, 1:1, 4:3, 3:4, and 21:9. Veo supports 16:9 and 9:16; the rest run on Seedance. Perfect for YouTube, TikTok, Reels, and Shorts.
Can I use references?
Yes, upload reference images for style or first-frame anchors, a reference video for camera movement and pacing, or reference audio to use as the soundtrack.
How is pricing determined?
It depends on the model. Some are flat-rate per tier; others are billed by usage, so longer, higher-resolution clips and added references cost a little more. On Auto a plain prompt is a flat Veo 3.1 Lite charge; runs that escalate to Seedance are quoted a ceiling up front and refunded down to what the run actually uses.
Can I choose the model?
Yes, pick one with the model selector on the form. By default a plain prompt runs a fast Veo 3.1 Lite draft, escalating to Seedance Pro only for references, 10s+, or Seedance-only ratios. Pin any Veo or Seedance tier to override.

Explore more pipelines

See all →
Image Generator
0.2-6.4
Image Generator →
Audio Generator
0.7-1.3
Audio Generator →
Music Generator
2.5-45.1
Music Generator →
Video Editor
13.1-29.8
Video Editor →