Questo articolo non è ancora disponibile nella tua lingua; mostriamo l'originale.
Blog

Guide

Happyhorse 1.0 Developer Guide: API Parameters, Pricing, and Integration

How to call the Happyhorse 1.0 video model, control cost per second, and pick sane defaults

Nov 16, 2026 1 min read
Cover art: faceted geometric prisms divided by straight lines on a deep pine-green ground, sage and mist-green facets, in a restrained style.
In questa pagina
A developer's workflow reimagined, generated with Happyhorse 1.0 — the video model this guide breaks down parameter by parameter.

What Happyhorse 1.0 Offers

Happyhorse 1.0 is a text-to-video model exposed as a single route, happyhorse-1.0. You send a text prompt and receive one video. This family has no image or audio input and no image-to-video or video-to-video route. Output length and resolution are request parameters, and both are also your cost levers because billing is per second of output video.

API Parameters

Only prompt is required. The other four fields have documented defaults. Omitting them is valid, but the resolution default is the higher-priced tier, so production requests should set the cost-bearing fields explicitly.

  • prompt (required, string, 1–5000 characters). A text prompt describing the video content, scene, and style.
  • aspect_ratio (optional, string, default 16:9). One of 16:9, 9:16, 1:1, 4:3, or 3:4. Send the enum as a string, not as a numeric ratio.
  • duration (optional, integer, default 5). Output length in seconds, from 3 to 15 inclusive. Longer-form work has to be assembled from multiple clips.
  • resolution (optional, string, default 1080p). One of 720p or 1080p. Use these lowercase enum values exactly.
  • seed (optional, integer, 0–2147483647). Fix the seed when you need a reproducible result while iterating on the prompt. Omit it when you want a new sample.

Official Per-Second Pricing

Happyhorse 1.0 bills per second of output video. The official rates are:

  • 720P: $0.168 per second
  • 1080P: $0.288 per second

A request costs duration × the rate for the chosen resolution. Because the default resolution is 1080p, a five-second request that omits resolution is billed at 5 × $0.288 = $1.44. The same five seconds at 720p is 5 × $0.168 = $0.84. A 15-second clip costs 15 × $0.288 = $4.32 at 1080P, or 15 × $0.168 = $2.52 at 720P.

Confirm the live panel on the happyhorse-1.0 model page if you need to re-check, and integrate against these official per-second tiers.

Integration Practices

Start with the cheapest official configuration that still lets you judge the prompt: 3 seconds at 720p, with a fixed seed. Once the prompt is right, promote that seed and prompt to the duration and resolution your product actually delivers. This avoids paying the 1080P rate while you are still iterating.

Write every field you care about explicitly. A missing resolution is not a “leave it to the model” choice — it selects 1080P at $0.288 per second. A missing duration selects 5 seconds. Persistent production configs should pin prompt, aspect ratio, duration, resolution, and seed (when reproducibility matters).

Validate client-side before you send: prompt length 1–5000 characters, aspect_ratio in the five-value enum, duration an integer from 3 to 15, resolution exactly 720p or 1080p, and seed an integer in range when present. Type mismatches and enum typos fail at validation; they are not prompt-quality problems.

If a request is rejected, inspect parameter validation first. If it succeeds but the result is off, confirm the five fields match what you intended — especially whether duration or resolution fell back to a default — then retry with the same seed before changing the prompt. If results are not reproducible, stop relying on omitted defaults and pin every field.

Where to Go Next

Start at 3 seconds and 720p with a fixed seed, prove the prompt, then promote the winning configuration to the resolution and duration your product actually delivers. The full route reference lives at happyhorse-1.0.

Inizia a creare con artefatti IA persistenti

Ottieni la tua chiave API e chiama modelli di immagini, video e audio tramite un'unica API basata sui task.

Leggi la documentazione API