AI writing assistant

Claude writes the words you do not want to: captions, titles, descriptions, hashtags and summaries. It writes text, and only text; no invented feature list here.

What it generates

Each skill produces one field of a draft, so you can take the caption and rewrite the hashtags, or the other way around.

Captions and titles

Platform-aware writing: a caption for a Reel, a title and description for a YouTube upload, sized to each platform’s limits.

Hashtags and summaries

Hashtag sets within each platform’s budget, and summaries that compress a source article into a post-sized brief.

Brand check

A reviewer that reads the draft against your own written guidelines and blocks the pipeline with its reasons when the draft misses them.

Your guidelines are the prompt

Every skill runs on an editable instruction file you own, per brand. Edit how your captions should sound, keep revision history, and restore any earlier version. The defaults are a starting point, not a personality you are stuck with.

Where generation fits the pipeline

Model
Claude, by Anthropic, called server-side only.
Budget
A per-workspace cap keeps generation spend bounded; the app tells you when it is reached.
Scope
Text generation only. ScheduleLater does not generate images, video or music.

Draft faster, publish the same

Generated drafts land in review like everything else. Nothing publishes itself.