Alt text writer and reviewer GPT that decides when an image needs a description and how long it should be
A custom GPT for content editors, designers and developers: upload an image with its surrounding context and it decides whether the image is informative, functional or decorative, then writes or reviews the alt text and any longer description it needs.
Agent in Design by AISkillHub Library
What it does
Good alt text depends on context: the same photo needs different alt text on a product page, in a news story, and as a link. This GPT asks for the context, applies the common decision process for alternative text, and writes text a screen reader user would find useful.
**What it does**
- Looks at an uploaded image (ChatGPT can read images you attach) together with the context you give: the page purpose, the text around the image, and whether the image is a link or button. - Classifies the image: **informative**, **functional** (link, button, icon), **decorative**, **complex** (chart, diagram, map), or **text in image**. - Writes alt text that conveys the purpose in context, avoids "image of", and does not repeat the caption. - For decorative images, recommends empty alt (`alt=""`) and explains why. - For charts and diagrams, writes a short alt plus a longer description or data summary to place on the page. - Reviews existing alt text you paste and explains what to change. - Outputs HTML snippets or CMS-ready text on request.
**When to use it**
- Publishing blog posts, product pages, newsletters and social posts - Accessibility remediation of an existing site section - Teaching a content team how to judge context
**What it will not do**
- It does not guess who people in a photo are. It describes visible features only when they matter in context, and names people only if you provide the names. - It does not infer sensitive attributes (ethnicity, religion, health, gender identity) from appearance. If the context makes something relevant, you supply it. - It is not a full accessibility audit and does not certify compliance with WCAG or any law.
**Example**
Image: line chart of monthly sign-ups. Context: blog post sentence "Sign-ups grew steadily after the March launch."
Output: Type complex. Alt: "Line chart of monthly sign-ups, January to December 2025, rising after March." Long description: a short paragraph plus a two-column table of the values read from the chart, marked "verify against source data" because values read off a chart image can be imprecise.
How to use it
**Set up as a ChatGPT custom GPT** (requires a ChatGPT plan that lets you create GPTs).
1. In ChatGPT, open the GPTs area from the sidebar (Explore GPTs) and choose **Create**. 2. Switch to the **Configure** tab instead of using the chat-based builder. 3. Set **Name** to `Alt Text Reviewer` and write a one-line **Description**. 4. Open `assistants/alt-text-reviewer/instructions.md`, copy everything, and paste it into the **Instructions** field. The file is written to fit inside the field's character limit. 5. Add **Conversation starters**: `Write alt text for this image. Context:`, `Review this alt text:`, `Is this image decorative?`, `Write a long description for this chart`. 6. Under **Knowledge**, upload (optional) your organization's accessibility or editorial guidelines, so the GPT follows house rules on length and style. 7. Under **Capabilities**: image generation and web search are not needed; turn them off. Keep Code Interpreter & Data Analysis off unless you want it to read chart data from an attached spreadsheet. 8. Test it in the preview pane on the right with a real example, then save and choose who can access it (only you, anyone with the link, or your workspace, depending on your plan).
**Using it:** attach the image and add one or two lines of context: where it appears, the nearby text or caption, and whether it is a link. Context changes the answer, so the GPT asks for it if you leave it out.
Files
Save each file at the path shown, relative to your project root.