Flocci Recall

Flocci Recall › Guides › Image memory

Content type

Image memory: how Recall handles images

How Flocci Recall saves and understands images: where they come from, what AI extracts, and the kind of fuzzy query that finds them again.

Direct answer

Recall treats images as a first-class content type with their own accent colour, their own filter chip and their own enrichment path. A direct image URL is filed as an image by fact, never by AI guess — the extension beats inference. Once a image is saved you can find it by describing it — the summary, key takeaways, topics, entities and bilingual keywords are all searchable.

Where images come from

Recall files a save as a image when it arrives from direct .jpg/.png/.webp/.avif/.heic links, Pinterest, photo URLs.

What Recall builds around a image

A direct image URL is filed as an image by fact, never by AI guess — the extension beats inference. Where a vision model is staged on the Flocci intelligence service, Recall describes what is actually in the frame and builds the card from that description.

Queries that find images again

These are the shapes of query Recall is built for — vague, human, and about meaning rather than metadata.

The honest bit

Vision degrades gracefully: if no vision provider is configured, the image still saves instantly with its metadata card — capture never blocks on AI.

Questions about images in Recall

How does Recall know something is a image?

From the URL and the content itself, before any AI runs. Image detection sits in a deliberately ordered signal list so more specific rules win — and where a fact is available, such as a file extension or the absence of a URL, the fact beats any AI inference.

What does Recall extract from a image?

A direct image URL is filed as an image by fact, never by AI guess — the extension beats inference. Where a vision model is staged on the Flocci intelligence service, Recall describes what is actually in the frame and builds the card from that description.

What can I search for to find a image again?

Describe it. Real queries that work include "the screenshot of the tax slab table", "that saree colour I liked", "the whiteboard photo from the architecture session". Recall interprets the description into concepts, ranks your whole library, and narrates what it found in one sentence.

Is anything about images handled differently?

Vision degrades gracefully: if no vision provider is configured, the image still saves instantly with its metadata card — capture never blocks on AI.

Related guides

Start saving things you will actually find again

One Google sign-in, free monthly credits, no card. If you already use any Flocci app, you are already signed in.

Open Flocci Recall