Runway Characters (live) vs Sume Avatar Video (async)

Pick Runway Characters when a viewer needs to talk to a face in real time, and Sume Avatar Video when you need a finished clip from a script. The Runway models guide lists a real-time model, gwm1_avatars ("Characters"), taking text conversation input and returning video and audio. Sume does not offer live sessions; it renders async jobs.

Side by side

How a Sume job runs

A request references a ready avatar with avatar_handle and provides exactly one of script or video_inputs. Sume estimates the video duration; scripts outside 4-60 seconds are rejected, so shorten them or split them into several jobs. quality is standard, plus (the default) or max.

Getting the result

The job is async. Poll GET /v1/jobs/{id}/status, read GET /v1/jobs/{id}/result, or pass mode: "webhook" with a webhook_url. Inline captions can be burned into the final MP4, and the docs offer a preview-first flow if you want to review first-frame stills before paying for a full render.

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: avatar-video-001" \
  -d '{
    "avatar_handle": "studio_presenter",
    "script": "Welcome back. Here is what changed this week.",
    "aspect_ratio": "16:9",
    "mode": "async"
  }'

Choosing

Live conversation means low-latency, turn-by-turn sessions, and the Sume docs do not describe that. If your product needs a presenter video, a product explainer or a localized greeting that can be queued and reviewed, the async route fits. If it needs a conversation, use a live product.