Synthesia Syren writes video as code: what Sume has instead

Syren is Synthesia's new prompt-driven video tool, and Sume does not have an equivalent. Sume has no agent that learns your brand from past videos and writes a whole motion-graphics video from a chat prompt. What Sume ships is narrower and more deterministic: a script-to-avatar API, first-frame previews, and a timeline that joins clips to one audio track.

Synthesia's own page describes Syren this way: "You describe the video you want, and Syren writes it as code and renders it live in front of you" (read 2026-10-11). It learns a brand kit from existing videos, takes prompts, documents, images or video links, and produces motion graphics, 3D, voiceover, music and avatars. Everything about Syren below comes from that page and its help article, both read 2026-10-11; everything about Sume comes from the linked Sume docs.

What Syren is, from Synthesia's own pages

The help article lists the steps: pick Wide 16:9 or Tall 9:16, describe the video, optionally attach documents, images, videos or PowerPoint files, press Send, watch a live preview, refine through chat, then generate at a chosen resolution. Videos can be up to 3 minutes long. Free videos on the Basic plan carry a watermark. Interactive avatars and live conversations are not available inside Syren.

The product page adds that Syren includes Synthesia's avatars and voices, but your own Personal Avatar or cloned voice requires building the video in Synthesia's regular editor instead.

What Sume offers for the same job

Sume's avatar route turns a ready avatar and a script into a talking video of 4 to 60 seconds. You can send one script, or ordered video_inputs with spoken scenes and silence beats, plus an optional scene prompt or photo and an optional product image (Generate avatar video). Before paying for the full render you can approve first-frame stills (Avatar video previews). To go past 60 seconds, you split the script and join the clips on Timeline 1.0, which takes one audio spine and up to 200 ordered video slots.

Nothing in those pages writes the video for you. You decide the script, the scenes and the cuts; Sume renders them. Avatar 1.0 is also English-only, so a Syren-style multilingual brand video is not something to expect from the avatar route.

Two more differences matter for planning. First, Sume prices avatar video per second before you submit, so a script's cost is known in advance: $0.184, $0.245 or $0.55 per second on the standard, plus and max tiers, with a small per-second surcharge when you add a product image. Syren's page quotes credits instead: a two-minute video with up to two rounds of edits typically uses 1,000 to 2,000 credits, and the pages I read do not convert credits to dollars. Second, Sume's output is a plain MP4 at 720p from the avatar job, and any design layer around it, such as cards or captions, is something you add with the caption option or a timeline.

Side by side

The table compares only what each vendor documents in the pages named in its caption.

How to choose this week

If you want a designed, motion-graphic explainer from a prompt and a pile of brand footage, Syren is the product built for that, and it is in early access with launch credits. If you want a repeatable pipeline where a script in a database becomes a 4 to 60 second English talking clip with a price known before you submit, use Sume's avatar route and a timeline for anything longer.

A practical split: let people prototype a brand look in Syren if they have access, then script the production runs through the API so each clip is reproducible with an idempotency key. For a decision about live avatars instead of clips, read the live versus rendered comparison.

One caution on timing: Syren is early access as of the pages read on 2026-10-11, and Synthesia's launch credits have an end date, so any workflow you build on it this week should be treated as a trial. Sume's avatar endpoints are generally available, which matters if a finance or legal reviewer needs a stable route to cite.