HappyHorse 1.1 on Cloudflare: 2,500-character prompt limit
Cloudflare's HappyHorse 1.1 text-to-video page caps the prompt at 2,500 characters and clips at 3 to 15 seconds. HappyHorse is not in the Sume catalog.

Cloudflare's HappyHorse 1.1 T2V page limits the prompt to 2,500 characters. It takes 3 to 15 second clips at 720P or 1080P. HappyHorse is not a model in the Sume video catalog, so there is no Sume prompt limit to compare it to.
What the Cloudflare page lists
| Field | Value |
|---|---|
| Duration | 3 to 15 s |
| Resolution | 720P, 1080P |
| Aspect ratios | 16:9, 9:16, 1:1, 4:3, 3:4 |
| Prompt | Max 2,500 characters |
| Seed | 0 to 2,147,483,647 |
| Watermark | Optional |
Why the limit matters
Prompts written for models with long structured briefs can pass 2,500 characters once you add shot lists, camera moves and dialogue. If your prompt generator can reach that length, truncate or split by shot before sending, and send one shot per request. 2,500 characters is roughly 400 words.
Note on conflicting resolutions
Alibaba's own model list describes happyhorse-1.1 text-to-video at 480P, 720P and 1080P. Cloudflare lists 720P and 1080P only. Treat the host you call as the authority for what is accepted.
On Sume
Sume's video catalog has Seedance, Wan 3.0, Kling 3, MiniMax H3, Gemini Omni Flash 1.1 and Grok Imagine; I found no HappyHorse id. Read GET /v1/videos/models for the live list before planning around any model.
Writing inside the cap
A good video prompt is rarely over 1,000 characters. When one is, it is usually because several shots are packed into one request. The better fix is one request per shot, with the same character and setting description repeated in each. Keep a short style paragraph, then one paragraph for action, one for camera and one for audio, and measure the length in code before the call.
A simple guard is to count characters in your template and fail the job locally when it passes 2,500, rather than paying for a request the host will reject. The same guard is cheap to keep when you later switch models, since each model has its own limit and Sume's GET /v1/videos/models is where you read the supported fields for the Sume ids.
Sources
Related posts
More in Models
- Ideogram 4.5 edits: first reference is the source, 4 more allowed
On Sume, the first Ideogram 4.5 input reference is the image being edited and up to 4 more are references: 5 total. Order them or the wrong photo gets edited.
- Ideogram 4.5 API cost: $0.0375 to $0.275 per image on Sume
Ideogram 4.5 via Sume is $0.0375, $0.075 or $0.275 per image (low, medium, high): $3.75, $7.50 or $27.50 per 100. Draft-then-upgrade: $6.50.
- Image-to-video prompt: write the motion, not the photo
When a photo is already pinned as the first frame, spend the prompt on motion, camera and sound, not on describing what the picture shows. A worked rewrite.
- Imagen 4 shut down Aug 17: Nano Banana 2.1 is the named replacement
Google lists gemini-nano-banana-2.1 as the Imagen 4 replacement. Imagen rows on Sume take no reference photos, while Nano Banana 2.1 takes up to 10.
Written by Sume