Hedra API gives 75+ models in one place vs Sume Avatar 1.0 route

Hedra's developer page advertises more than 75 models from 13 providers behind a unified api.hedra.com, and says jobs return 202. Sume Avatar 1.0 is narrower: two routes, one for avatar creation and one for talking video. Choose breadth if you want to switch between many generation models; choose Sume if you want a stable avatar workflow with preview, captions and spend controls.

What Hedra lists

The page presents Hedra as a single integration point for many models. Details of which models support avatar talking video are on the vendor's page, so check there.

What the Sume route gives you

Avatar 1.0 pairs an avatar handle with a script, optional product image and scene, inline captions and previews. Async jobs also return 202. Reads are GET /v1/avatar-videos, GET /v1/avatar-videos/:id and GET /v1/avatar-1.0/avatars, per the docs.

Captions are inline: captions with enabled, style and language. A caption failure soft-fails, so the video still succeeds with captions.status=failed.

Choosing

If you are experimenting, a broad catalog saves integration work. If you are shipping a repeated format, a narrow route reduces surface for bugs and makes cost predictable.

Either way, keep your own script, avatar handle and webhook handler vendor-neutral so you can switch later.