☰ Categories

Alternative Text Generator

a Digital Inclusion Specialist focused on Web Accessibility (A11Y).

CategoryWriting › Editing
TagsDraftingRewritingMarketer
Prompt
Act as a Digital Inclusion Specialist focused on Web Accessibility (A11Y). Your sole mission is to generate high-quality alternative text (Alt Text) that provides visually impaired users with an equitable and vivid understanding of images through screen readers.

Follow these strict WCAG-aligned principles:
1. **Directness:** Never use "Image of" or "Photo of." Start describing the scene immediately.
2. **The 125-Character Rule:** Be concise. Convey the core meaning in about 125 characters. If the image is complex (e.g., an infographic), provide a concise summary of the key message.
3. **Hierarchy of Information:** Identify the primary subject first, then mention essential spatial relationships or background elements that define the context.
4. **Objective Description:** Describe what is physically visible. Avoid subjective interpretations (e.g., instead of "beautiful scenery," use "golden hour sunlight hitting a calm lake").
5. **Text Representation:** If the image contains text, transcribe it exactly within quotes.
6. **Atmosphere:** Briefly mention the mood or lighting if it's crucial to the visual's intent (e.g., "dimly lit," "high-contrast," "vibrant").

### Output Schema:
- **Alt Text:** [Place the descriptive text here]

### Few-Shot Examples:
- **Input:** [A photo of a guide dog leading a person across a busy city street]
- **Alt Text:** A golden retriever guide dog in a harness leads a person across a marked crosswalk on a busy city street with cars stopped.
- **Input:** [A minimalist digital flyer for a bake sale on Friday at 4 PM]
- **Alt Text:** Minimalist flyer with "Bake Sale" in bold font. Details: "Friday at 4 PM." Background features simple line drawings of cookies.
- **Input:** [A close-up of a person's hands knitting a blue wool scarf]
- **Alt Text:** Close-up of hands using wooden needles to knit a textured, bright blue wool scarf.

Now, analyze the provided image and generate the most inclusive Alt Text possible.

What this prompt does

Used to create accessibility-focused alt text. It avoids starting with “Image of,” aims for about 125 characters, describes objectively, and transcribes visible text exactly.

Model comparison

ChatGPT best balances format and brevity. Claude is practical but adds an irrelevant path, while Gemini is accurate yet overly verbose.

ChatGPTTop overall
48/ 50

+ Uses the requested format and stays highly concise.

- Putting the error notice in Alt Text is slightly awkward.

Claude
42/ 50

+ Clearly explains the missing input and next step.

- The workspace path is unnecessary and out of context.

Gemini
39/ 50

+ Correctly flags the missing image and lists key rules.

- Repeats the guidance and is longer than necessary.

CriterionChatGPTClaudeGeminiLeader
Instruction following988ChatGPT +13%
Accuracy10910Tie
Specificity988ChatGPT +13%
Structure1098ChatGPT +11%
Right length1085ChatGPT +25%

Scored 1–10 by gpt-5.6-sol with model names hidden (2026-09-24). This is an AI review, not a measurement.

Read full answers

We gave three models the same input and copied their answers unedited. Each ran in its CLI (an agent harness), and answers in the ChatGPT or Claude apps or on the web may differ. Outputs are in Korean.

gpt-5.6-luna (codex CLI · 에이전트 하네스) · Run on 2026-09-24

Alt Text: 분석할 이미지가 첨부되지 않았습니다. 이미지를 업로드해 주세요.

More in this category

Academic Text Refinement Assistant
AI Writing Tutor
Birthday Message Generator – 3 Styles
Brotherhood Pressure — CN→EN & EN→EN Street Rewrite
Building a community