Generate alt text and SEO captions for an image with Gemini Vision

Hand-writing accessibility alt text for twenty product photos is slow, repetitive, and easy to skip. Skipping it is also a measurable accessibility regression: screen reader users hit "image" announcements with no context, and the SERP loses every long-tail keyword the photo could have anchored. This tool replaces the manual pass with a single Gemini Vision (US) call: drop the image, get back a JSON object with alt text, a 1-2 sentence caption, a kebab-case SEO filename, three to five keyword tags, and the dominant subject phrase.

The trade is honest. Image bytes are sent to Google Gemini Vision (United States); the bytes do not stay on your machine. F-13 carries an explicit consent gate: the first time you submit an image while logged in, the disclosure modal appears and the submission is held until you accept. You can withdraw consent on /account, which flips the column back and forces the modal again on the next attempt. The privacy badge on this page reads external-api so you can eyeball the data flow before you read the prose.

Each field is editable after the response renders, so the AI output is a draft, not a verdict. Alt text is capped at 125 characters and truncated at a word boundary if Gemini ignores the maximum. The credit is refunded automatically if the JSON parse fails (worker retries twice; on final failure, refund) or if the safety filter rejects the image. You never pay for our integration mistakes.