Post alt OCR

I use Phanpy for all my Fediverse now, with two exceptions (a) admin/mod work and (b) posting.

(a) isn't Phanpy's problem, masto client is OK.

But when posting I often include screenshots, and I never post an image without alt text, and the native client has the "detect text from picture" feature that makes it easy. So when I need that I switch back to masto client.

(Not a matter of life & death, obviously, but something useful for the roadmap.)

Start by tracing Phanpy’s posting flow, especially screenshot attachment and alt-text handling. Compare that flow with the requested native-client text-detection behavior, then define what OCR support should produce and when it is complete. The issue names no files or tests, so locating the relevant posting entry points is part of the initial investigation.

Written by the indexing model from the issue text.