HeyGen POST /v3/templates makes a template from video; Sume has none

HeyGen's October 2026 changelog entry adds POST /v3/templates, which creates a private template from an existing generated video. Sume has no template resource for avatar video. What you reuse is the ready avatar, referenced by avatar_handle, and the request body that you store yourself.

The HeyGen template calls

From the HeyGen API changelog, read 2026-10-02:

What you reuse on Sume

The reusable piece is the avatar. Create it once with POST /v1/avatar-1.0/generate, then send its handle on every avatar video request. A request carries a script or video_inputs, an optional scene, quality, aspect_ratio and captions.

If you want to review before paying for a full render, avatar video previews create first-frame stills you can regenerate, and generate-video then starts the final render from that preview.

Making your own template

Treat the JSON body as the template. Store it with placeholders for the script and the scene, fill them per recipient, and submit. That gives you version control in your own repository instead of an edit version in a vendor API.

If your repeat unit is bigger than one clip, a Format is Sume's saved procedure; the bulk runs page documents queuing up to 100 Format runs with a concurrency window. That is separate from the avatar route and is a different kind of reuse.

• No expected_edit_version: there is nothing server-side to overwrite.
• Pin the avatar handle in your stored body so every render uses the same presenter.

Trade-off

HeyGen lets people edit the template composition in its product and then fill it by API. Sume asks you to own the request. That is less editable by non-developers and more transparent to developers.