Narakeet Markdown-to-video API vs Sume Timeline render

Narakeet generates the voice for you from a Markdown script; Sume Timeline does not. Timeline 1.0 takes one audio spine you already have plus up to 200 ordered video slots and returns one MP4, with no provider inference behind it.

So the two are not rivals for the same step. Narakeet's pages (read 2026-10-10) describe converting scripts, slides and Markdown into narrated video through an API. Sume's Timeline 1.0 docs describe an assembly job, and a voice comes from a separate Sume TTS request.

What each API takes in

Narakeet lists four API endpoints: text to speech, Markdown to video, a voice list and an account credits call. Its video API only supports Markdown scripts. A Sume Timeline request is a JSON document with an audio block and a video[] list, and every URL must already be a workspace media artifact.

A fair cost example

A 3-minute narrated video makes a rough comparison. At Narakeet's smallest package rate of about $0.20 per minute, it is about $0.60. On Sume, Timeline reserves ceil(180 / 60) = 3 output minutes at $0.10, so $0.30, but that excludes the TTS request that makes the audio and the clips that fill the slots.

Do not read $0.30 against $0.60 as a verdict. Narakeet's price includes the voice and the slide-to-video generation; Sume's number is only the assembly. The comparison only makes sense when you add the Sume TTS and generation lines for your case.

Where Sume Timeline is the better fit

Choose Timeline when the voice or footage comes from elsewhere. It supports fit modes (cover, contain, stretch, blur), transitions such as fade and dissolve on slots after the first, an optional soundtrack with ducking, and an output size of 256 to 2160 on each edge. The default is 1080 x 1920, which suits vertical posts.

The plan endpoint is the feature Narakeet's pages do not mention: it compiles your document and returns billable_minutes and estimated_cost_usd_micros without creating a job. For a pipeline that must stay inside a budget, an unbilled dry run is useful.

Choose Narakeet when the input is a slide deck or Markdown and you want one call to give you the finished narration.

Combining them

Nothing stops a pipeline from using both, with one caveat I have not verified: whether files that Narakeet produces can be imported as Timeline sources. Sume's docs say inputs must be workspace media.sume.com artifacts and tell you to import first. Read the media import path in the API reference before relying on it.