Last Updated: September 2026

Media analysis lets Capacities understand your media files so you can search and work with them like any other object in your space.

It only runs in spaces where AI is turned on. You start it yourself (for example by clicking Analyze). If AI is off for a space, none of its media can be analyzed.

Availability

Analysis uses your AI budget. Longer or larger files use more budget.

What analysis does

• Turns images, weblinks, PDFs, and audio into searchable information on the object.
• Makes that information available as context for the AI assistant, so the model does not need to send the original file again and again.
• For images, can suggest a better title when the current title is generic (for example “image” or “screenshot”), and leaves meaningful titles you wrote yourself unchanged.

Image analysis

Image analysis extracts a better title, description, text (OCR), category, and colors, and stores them on the image object.

Extracted text and color swatches appear in the image toolbox and embed UI, so you can review the results while you keep working in your notes.

For full details, see Image analysis.

Weblink analysis

Weblink analysis runs on desktop. It adds a summary, keywords, topic, category, and saved page content to the weblink object so you can search and filter by the page, and use it as context in AI chat.

For full details, see Weblink analysis.

PDF analysis

PDF analysis adds a page-by-page Reader view, a short summary, keywords, document type, and topic to the PDF object so you can:

• Search inside the document and jump to the matching page.
• Filter PDFs by document type or topic, and use the extracted information as context for the AI assistant.

For full details, see PDF analysis.

Pro and Believer users can also highlight passages, add notes, and capture snapshots without running analysis. Annotations do not need AI.

Audio analysis

Audio analysis is in early access for Believer. After early access, it will be part of Capacities Pro.

It adds a transcript and a summary to any audio object so you can search the spoken words and use them as context for the AI assistant. That includes recordings you make in Capacities, voice notes from WhatsApp or Telegram, and other supported audio files in your space.

For full details, see Audio analysis.