Sources

h3 max source notes

These notes separate current public facts from implementation choices on h3maxapi.com. Provider docs and pricing can change, so production integrations should re-check the linked pages before shipping a paid workflow.

Primary API references

fal H3 Max text-to-video API reference lists the hosted text route, input fields, async queue examples, and server-side credential guidance.
fal H3 Max image-to-video API reference lists the hosted image route, image URL fields, resolution options, and queue payload shape.
fal queue documentation describes submit, status, and result calls used by asynchronous generation jobs.
fal key-based authentication documentation explains provider-side API keys. h3maxapi.com keeps that provider key on the server.

Model and launch context

fal's H3 Max launch article presents H3 Max as a post-trained H3 variant and includes fal's quality and speed claims.
MiniMax H3 launch article describes MiniMax H3 as a multimodal model family with text, image, video, and audio context.
MiniMax H3 on Hugging Face provides model-card context, community activity, and public release references.
MiniMax H3 GitHub repository gives a community-visible engineering checkpoint for examples, issues, and release material.

How these sources shaped the homepage

The homepage uses site-issued keys because provider docs warn against putting provider credentials in client-side code.
The workbench includes both text and image modes because those are the two high-intent H3 Max API references surfaced for the keyword.
The route list mirrors the first useful product loop: create a key, manage it, submit a generation, and poll a generation.
The copy avoids fixed pricing claims because public pricing surfaces can change by promotion and model route.