How to make an AI avatar video longer than 60 seconds

To make an AI avatar video longer than 60 seconds with Sume, split the script into parts Sume estimates at 4-60 seconds each, render every part with POST /v1/avatar-1.0/talking-video, then join the parts with POST /v1/timeline-1.0/render. The 60-second window applies to each avatar job; a Timeline render can run up to 1,800 seconds.

The steps below come from Sume's Generate avatar video, Timeline 1.0, and Audio detach docs, read on 2026-09-26. The single-job request itself is covered in Talking avatar video API.

How should I split and render the parts?

Sume estimates each job's length from its script, and the docs' advice for material past the window is to split it into multiple jobs. Cut the script at sentence or paragraph breaks so each part lands comfortably inside 4-60 seconds, then submit one talking-video job per part:

• Send the same avatar_handle, scene, aspect_ratio, and quality on every part. Each part is still its own job, so review the finished parts together before you assemble them.
• Give each part its own Idempotency-Key, and poll each job. Completed results can include public media.sume.com video artifacts, which is the kind of URL Timeline requires.

curl -X POST https://api.sume.com/v1/avatar-1.0/talking-video \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: onboarding-course-part-01" \
  -d '{
    "avatar_handle": "product_host",
    "script": "Welcome to the Acme onboarding course. First, we set up your workspace.",
    "aspect_ratio": "9:16",
    "quality": "plus"
  }'

How do I join the parts and keep their audio?

Timeline builds one MP4 on one audio spine plus ordered video[] slots, and the finished file's audio comes from that spine, so the spine has to carry the parts' own audio. Detach each part's audio with POST /v1/audio-detach: its default output is a sample-exact wav, the format Timeline's audio.url wants, and the result reports duration_seconds. Then build the render:

• audio.parts[]: the part wavs in speaking order, up to 20 gapless slices joined in the sample domain with no re-synthesis. Set audio.duration_seconds to their total; if declared part lengths sum to less, the render refuses with audio_parts_shorter_than_duration.
• video[]: one slot per part, with the part's MP4 as source_url, its length as duration, and the running total of earlier parts as start. video[0].start must be 0, and coverage may trail the spine by at most 0.5 seconds.
• Optional: a transition on slots after the first (fade, dissolve, wipeleft, and others, 1 second at most), and output.width / output.height. Timeline's default output is 1080×1920, while Avatar Video's resolution is currently 720p.
• To reuse the joined audio elsewhere, Timeline audio concat returns one file plus segments[] offsets to use as video[].start.

curl -X POST https://api.sume.com/v1/timeline-1.0/render \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: onboarding-course-assemble-01" \
  -d '{
    "audio": {
      "duration_seconds": 104,
      "parts": [
        { "url": "https://media.sume.com/artifacts/artf_demo/part1.wav" },
        { "url": "https://media.sume.com/artifacts/artf_demo/part2.wav" }
      ]
    },
    "video": [
      { "source_url": "https://media.sume.com/artifacts/artf_demo/part1.mp4", "start": 0, "duration": 58 },
      { "source_url": "https://media.sume.com/artifacts/artf_demo/part2.mp4", "start": 58, "duration": 46 }
    ]
  }'

What are the limits?

Run POST /v1/timeline-1.0/plan first: it is an unbilled compile preflight that returns the duration, segment count, and estimated cost without creating a job.

What does a longer avatar video cost?

Each part bills at the avatar-video rate for its length and tier: $0.184/s standard, $0.245/s plus, $0.55/s max (no product image). Each detach is a flat per-job rate listed on the Audio detach page. The render is listed at $0.10 per output minute on API pricing, reserved as the spine's length rounded up to whole minutes. Each of these is plus a 5.5% agent fee by default.