Draft captions that get found and get replies - scored for SEO and engagement against each platform's numbers. No account, nothing stored.
1. Photos 1-8 images, jpg / png / webp, max 15 MB each
Drag photos here or
Best results with several shots of the same thing - that's what this is for.
Everything runs in this tab. Photos are decoded to a canvas and measured locally - they are not uploaded anywhere
unless you switch on the optional caption writer below.
2. Where are you posting?
The primary keyword is what the SEO score is measured against - whether it appears before the "see more" fold,
how often it repeats, and whether the alt text carries it. Leave it blank and those checks are skipped rather than guessed.
3. Have a model draft them optional
The scoring is always on and needs no key - it grades whatever is in the caption box, whether a model wrote it or
you did. What a key buys you is the first draft. Leave this off and you get editable scaffolds instead: the right
structure with the keyword already placed, for you to fill in. A scaffold you complete beats a caption a model invented.
Provider
A relay is about twenty lines that forwards the request and adds the CORS headers the provider omits. Your key still
travels from this page and is never stored by the relay. If you have not set one up, use Anthropic instead.
Whichever you pick, the key stays in this tab's memory only - never storage, never anywhere but that provider.
Ticking this box is the only thing that sends your photos off the machine; untick it and they never leave.
This report is not saved. It lives in this tab's memory only. Reloading or closing destroys it.
Copy what you need or use Export Markdown first.