A photo of a receipt, an invoice, a shelf label, a screenshot or a business card carries words that describe it better than anything else. On a plan with vector search Opensolr reads those words, and your search finds them.
- The receipt is found by the shop's name, the total or the receipt number printed on it.
- The invoice by the company on it, the label by its product code, the screenshot by what it says.
- The printed text weighs the most of everything a search looks at, so a receipt is found by what is written on it, not only by looking like a receipt.
There is nothing to turn on. The text lands in your index next to everything else, and the search you already use looks through it. The OCR switch in filters keeps only the photos that have printed text.
- Every photo sent to Opensolr is read for printed text on Opensolr's servers, while the same photos are read into words. Nothing is read on your phone.
- A reading is kept only when it holds real words. The few stray letters found on a holiday photo are dropped, so such a photo carries no printed text.
- The text is searchable but never shown as a description of the photo: what the photo shows stays separate. You find it in the photo's details, under Text printed in the photo.
- It is not offered by autocomplete, so a receipt does not fill the suggestions with its numbers.
Nothing extra. A photo costs a tenth of an AI request whether Opensolr read its words, its printed text, or both. A picture read once is never read again: the reading is kept against the picture itself.
The reading also survives a later pass that cannot read the photo, such as a plan without vector search or a month whose AI requests are used up, as long as it is the same file. The md5 of the file proves that.
On a plan without vector search no photo is read, so there is no printed text to search. See without photo recognition.
Documents, on the Sync screen, has the photos with printed text read again. See Re-read, Documents and Reset. The same reading is part of Opensolr's site search too: text in scans and pictures.