Happy Horse Reference AI Generator
Generated on PixelDojo with Happy Horse Reference. Produced by PixelDojo's generation pipeline.
Happy Horse Reference costs 12 credits for a 3 second 720p clip on the 1.0 engine, and our run on August 26, 2026 took about 4 minutes from request to finished file. The mode takes one or more reference images plus a prompt and drives the subject through motion, rather than starting from a blank scene like the text-to-video mode or animating a single still like the image-to-video mode. This page covers the real charge, the measured turnaround, what the clip actually contains, and when to reach for a sibling mode instead.
The 12 credit reference clip, first result kept
Every example below was produced on PixelDojo. Hover to see the prompt.
Slow cinematic camera orbit around the earbuds case, soft studio lighting shifts, the lid opens gently to reveal the earbuds, product commercial style
Happy Horse Reference
What the reference mode actually does
One image becomes a scene
We sent the earbuds product photo as the sole reference and the exact orbit prompt used for the image-to-video test. The model treated the photo as the subject to carry through motion rather than as a first frame to extend.
Two engine versions, two prices
The tool defaults to version 1.1, priced at 3 credits per second at 720p and 4 at 1080p. Version 1.0, which this run used to match the catalog listing above, charges 4 and 6. Passing version explicitly is the only way to control which one you get.
1 to 9 reference images
The schema accepts up to nine reference URLs in one request. Prompts can bind specific subjects to specific images using character1, character2 tokens when you send more than one.
Duration floor is 3 seconds, not 2
The reference mode's own schema already documents a 3 to 15 second range, unlike the plain text and image modes in the same family whose schemas allow 2 seconds but reject it upstream. 3 seconds is the real floor everywhere in this family.
Five aspect ratios
16:9, 9:16, 1:1, 4:3, and 3:4 are all available on the aspectRatio field. We left it at the 16:9 default for this run.
What came back
A 1280 by 720, 24 frames per second H.264 file, about 3 seconds of playable video, 1.1 megabytes. Measured with ffprobe against the file itself, not estimated.
One reference image, one prompt, one run on August 26, 2026: 12 credits, about 4 minutes, a 1280 by 720 clip.
Why Choose Pixel Dojo for Happy Horse Reference
Professional-quality results with cutting-edge AI technology
12 credits for the 1.0 engine
4 credits per second at 720p, 3 seconds minimum. The cheaper 1.1 engine runs the same clip for 9 credits if you pass version 1.1.
About 4 minutes, measured
251 seconds from submission to a finished file on August 26, 2026, with no retries.
Subject identity from a single photo
No character sheet or multi-shot reference set required. One product photo and a prompt were enough to drive the reference mode.
How It Works
How we ran it, so you can repeat it:
Send one reference image
We used the same product still from our image-to-video test as the single entry in reference_urls.
Pick a version and duration on purpose
version 1.0, duration 3, resolution 720p. Skipping version defaults you to the cheaper 1.1 engine instead.
Run it in the app or over the API
Generate from the Happy Horse page's reference mode, or POST to /api/v1/models/happyhorse-1.0-r2v/run. Ours took about 4 minutes and charged 12 credits.
Loved by creators on PixelDojo
Real feedback from people using PixelDojo, pulled from our in-product surveys.
Because it is awesome
exceptional quality and great overall design of platform and interface. very intuative. love the creative freedom.
Qwen image 2 is amazing!!
Creative freedom, range of tools and options.
I love the training feature
the quality is the best
Explore more AI tools on PixelDojo
AI Tools
Compare & Switch
- Best AI Image Generators
- Best AI Video Generators
- Midjourney Alternatives
- Civitai Alternatives
- Runway Alternatives
- Leonardo Alternatives
- Pika Alternatives
- Luma Alternatives
- Magnific Alternatives
- Veo Alternatives
- Flux Alternatives
- Freepik Alternatives
- Seedance Alternatives
- Seedream Alternatives
- Pixverse Alternatives
- GPT Image Alternatives
- Synthesia Alternatives
- Playground Alternatives
- NightCafe Alternatives
- Canva AI Alternatives
- ElevenLabs Alternatives
- ComfyUI Alternatives
- Fal Alternatives
- Replicate Alternatives
Common Questions
Everything you need to know about Happy Horse Reference
What does Happy Horse Reference cost?
4 credits per second at 720p and 6 at 1080p on the 1.0 engine, with a 3 second minimum, so the cheapest clip is 12 credits. Version 1.1 is cheaper: 3 credits per second at 720p, 9 credits at the 3 second minimum.
How fast is Happy Horse Reference?
About 4 minutes in our August 26, 2026 run, measured from submission to the finished file. Alibaba's video pipeline is asynchronous, so turnaround moves with queue load.
How do I call it from the API?
POST to /api/v1/models/happyhorse-1.0-r2v/run with your API key, a prompt, and a reference_urls array of one to nine image URLs. Add version, resolution, aspectRatio, and duration to control price.
How is Reference different from Image-to-Video?
Image-to-Video treats your photo as a first frame and animates forward from it. Reference treats the photo as a subject to carry through a scene the prompt describes, and accepts up to nine reference images instead of one.
How is Reference different from Video Edit?
Reference generates a new clip from images. Video Edit takes a video you already have and restyles or locally replaces something in it, keeping the original motion.
Does the duration really start at 3 seconds?
Yes. The reference mode's schema documents 3 to 15 seconds directly, and in our testing the sibling text and image modes reject anything under 3 seconds too, even though their own schemas technically allow 2.