falImage to videoAvailable now

fal-ai/kling-video/v4/reference-to-video

fal-ai/kling-video/v4/reference-to-video is an image to video tool from fal that runs in the background and returns video. Pay as you go from $0.1222 per second, with one API key.

Try itfrom $0.1222 / second

Pricing

resolution: 720p $0.2041, 1080p (default) $0.3055 per second. Billed on duration, default 5. With reference videos, a lower per-second rate applies, their seconds count too, and at least twice the output is billed. The final charge comes from usage.

Limits

Runs: In the background
Returns: video
Longest queue wait: 3600s
Run deadline: 3600s
Largest inline input: 1024 KB
Largest input file: 500 MB
Largest output: 4768 MB
Outputs per run: 16

Input

Field	Kind	Default	Description
`prompt`	text	—	Text prompt. Reference inputs as @Image1, @Video1, @Element1, etc. Either prompt or multi_prompt must be provided.
`multi_prompt`	JSON	—	Per-shot prompts for multi-shot generation. The video duration is the sum of the shot durations.
`duration`	number (whole) 3 to 30	5	Video duration in seconds (3-30).
`start_image_url`	URL or file	—	Optional image to use as the first frame. Counts toward the image reference limit.
`middle_image_urls`	URLs or files	—	Optional ordered keyframes between the first and last frames (max 8). Count toward the image reference limit.
`end_image_url`	URL or file	—	Optional image to use as the last frame. Counts toward the image reference limit.
`image_urls`	URLs or files	—	Reference images, referenced as @Image1, @Image2, ... (max 10).
`video_urls`	URLs or files	—	Reference videos (3-30s each, 30s combined with element videos), referenced as @Video1, @Video2, ... (max 5).
`elements`	JSON	—	Characters/objects, referenced as @Element1, @Element2, ... (max 7). Up to 3 elements may carry a voice_id and up to 3 may be video elements.
`voice_ids`	list	—	Voices referenced in the prompt as @Voice1, @Voice2, ... (max 3, including voices bound to elements). Create one with https://fal.ai/models/fal-ai/kling-video/create-voice
`resolution`	one of 720p, 1080p	1080p	Output resolution.
`aspect_ratio`	one of auto, 16:9, 9:16, 1:1, 21:9	16:9	Aspect ratio of the generated video. 'auto' adapts it to the references.

Same request in code

curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "fal",
    "model": "fal-ai/kling-video/v4/reference-to-video",
    "input": {}
  }'