Skip to main content

Media library

The media library (sidebar → Library) is your central store for images, videos, and documents. Upload once, reuse across any number of posts, and organize into collections that agents can draw from.

Two tabs: Media (everything, filterable by All / Images / Videos, paged as you scroll) and Collections. A storage meter in the header shows used vs. plan quota.

Upload​

Drop files (or browse): images, video, and PDF documents. Details worth knowing:

  • HEIC/HEIF photos are converted to JPEG before they leave your browser — the stored file is the converted one.
  • Uploads use a presigned URL flow: the browser PUTs directly to object storage (R2) and then registers the record. Videos are served back with full byte-range support, so they stream and seek correctly.
  • The effective size ceiling is the smaller of your plan's per-file limit and the absolute ceiling:
TypeAbsolute ceilingFree-plan ceiling
Images25 MB10 MB
Video100 MB25 MB
Documents100 MB25 MB
  • Allowed types: JPEG, PNG, WebP, GIF, HEIC/HEIF; MP4, WebM, QuickTime; PDF (stored as documents — used by LinkedIn document posts).
  • Storage quota — your plan's total storage (5 GB on Starter up to 500 GB on Agency; 1 GB on the free bucket). Exceeding it blocks the upload with a clear message; free space by deleting unused media.

Per-platform limits are often lower still — and the smaller of the two is enforced when media is attached to a post. See platform limits.

AI analysis​

Every uploaded image and video is analyzed automatically:

  • Images — one vision pass produces a 2–3 sentence description (subjects, setting, colors, text, actions), brand-context aware. This is what lets captions "see" your media.
  • Videos — an extraction container pulls four evenly-spaced frames and up to 60 seconds of audio; each frame is described, the audio is transcribed, and a synthesis pass writes the final description. Frame zero becomes the video's poster/thumbnail.
  • Analysis is metered as AI actions (a video costs its frames plus transcription and synthesis) — if today's remaining allowance can't cover a video, the job waits and tells you when quota resets. If analysis is slow or you change the content, you can edit a description or force a re-run from the media detail panel.

Thumbnails are generated server-side and are not AI — no allowance consumed.

Media detail​

Click any item for the full record: preview (video with controls), type, size, dimensions, upload date, and its AI description. You'll also find:

  • Posted in — every post using this media, with status and time, linking to the post.
  • Add to collection / Remove from collection.
  • Edit (images only) — opens the crop editor.
  • Download and Delete.

Delete guard — media attached to any draft, scheduled, or in-flight post can't be deleted, nor media on a post with an active recurrence. The error names the blocking posts. Published history and collection memberships don't block deletion.

Collections​

A collection is an album with a name and an optional description (the description is used as AI context when agents pick from it). Collections are how you tell agents what material is approved:

  • Create from the library header; add media from any media detail, or in bulk from a collection's Add Media picker (already-added items are hidden).
  • Reorder with drag handles — the position is the order agents consume media in.
  • Attach collections to an agent as its source pool (see Agents).

Using media in posts​

When you attach media in the composer, each file is validated against the destination platform's constraints for that post type — a video that's fine as a YouTube long-form upload (≤4 h) will be flagged if you also target it at a YouTube Short (≤3 min). Media keeps its sortOrder, which matters for Instagram carousels and Reddit galleries.