BAbannerbear

Generate AI video

Changed on

Generate a video from a prompt or image with AI. Async — returns 202 with a pending job; poll GET /tool_jobs/{uid} until status is completed or failed.

post/tools/generate_ai_video

Request

  • Base URL
    https://api.bannerbear.com/v5
  • Auth
    HTTP bearer

Request body

promptstring required

What happens in the clip — describe the subject, motion, camera and style

model'seedance_1_lite' | 'seedance_2' | 'kling_2_5_turbo_pro' | 'veo_3_1_fast' required

AI credit cost varies by model, length and resolution

durationstring

Each model offers some of these lengths — Kling 5 or 10, Veo 4, 6 or 8, Seedance any

resolution'480p' | '720p' | '1080p'

Higher resolutions cost more per second — Kling's is fixed

aspect_ratio'16:9' | '9:16' | '1:1' | '4:3' | '3:4'

Ignored when a starting image is given — the clip takes the image's shape

image_urlstring uri

Optional — the clip begins from this frame and animates it

audio'on' | 'off'

Generated sound, on the models that can — Veo charges less without it

metadatastring

Arbitrary metadata string stored with the tool run

Response

Tool run accepted

uidstring
tool'generate_ai_video'
status'pending' | 'running' | 'completed' | 'failed'

pending → running → completed / failed

progressinteger

0-100 job progress, 100 once completed

inputsobject

Inputs the job was created with

metadatastring nullable

Arbitrary metadata string round-tripped from the create request

selfstring uri
created_atstring date-time
completed_atstring date-time nullable
error_messagestring nullable

Human-readable failure reason (only present on failed jobs)

Changes