Faceless education video episodes, 1-5 minutes, on Sume Timeline

A 1-5 minute faceless education episode fits one Timeline 1.0 render: a single narration file as the audio spine and an ordered list of scene stills or clips as video[]. Timeline accepts up to 1800 seconds of audio and 200 video slots, so five minutes is well inside the limits. Reuse the same document shape each episode and only swap the files.

Higgsfield's changelog (Aug 20, 2026) describes Faceless Studio as a standalone product with Education, History, Kids and Storytelling themes, 1-5 minute lengths, and a saved channel style and voice. Sume has no channel object; the repeatable part is your own template. Sume facts are from Timeline 1.0, read 2026-10-01.

What does an episode look like as a Timeline document?

One audio spine with duration_seconds, plus one video[] slot per scene. Each slot has source_url, start, and duration; the first start must be 0 and later starts must increase. Stills are static holds, and fit can be contain if you do not want a crop.

How do I keep the same style every episode?

Keep a fixed recipe on your side: the same output.width and output.height, the same fit, the same transition type, and the same narrator voice for the audio file. Generate scene stills from one saved prompt prefix so they look alike. Sume does not store a channel style for you, so the template lives in your code or config.

How do I check the cost before rendering?

Call POST /v1/timeline-1.0/plan. It is unbilled, creates no job, and returns billable_minutes and estimated_cost_usd_micros. The docs say the reserve for a render is ceil(audio.duration_seconds / 60) minutes. See Timeline billable minutes for the rounding.

What does Sume not do here?

Timeline assembles files you already have. It does not write the script, pick a theme, or animate characters on its own. For the narration and b-roll steps, read faceless video API: voiceover, b-roll, music.