Image alt text draft
Draft image alternative text for human review.
Image alt text draft
Drop your files here
or choose files from your device
How to use Image alt text draft
Image alt text draft produces a first draft of alternative text for an image with a local ViT-GPT2 captioning model. The draft is a short scene description that you edit to add purpose, context and any important visible text before placing it in an alt attribute. It can be copied or downloaded as alt-text.txt.
- Choose a source file. If the tool accepts multiple files, arrange them in the order you want.
- Check the source format before running the operation.
- Run image alt text draft, review the output, then use the available copy or download controls.
What this tool supports
Downloads an inference model on first run. Chat and generation require WebGPU; smaller recognition models use browser WebAssembly. AI output can omit or invent details. Review it before use.
Limits and processing
Text sources are capped at 8,000 characters; image search accepts up to 20 images and a 500-character query. Models may truncate long text. First downloads can be hundreds of MB and need browser memory.
Example source
A useful tool makes everyday work easier. Review the results before sharing.
Frequently asked questions
How do I write alt text with AI?
Upload the image and run the tool to get a short description from a local model. Then rewrite it for the page: say why the image is there, include any important text it shows, and remove details that are wrong.
Can I use the AI alt text without editing it?
It is meant for human review. The tool warns that the model can omit or invent details, and it knows nothing about the surrounding page, so check accuracy and add context before publishing.
Can I generate alt text for several images at once?
Only the first selected image is described in each run. Process images one at a time and copy or download each draft before moving to the next.
Where is my input processed?
Local mode downloads public model/runtime assets and keeps input in this browser. Models may remain cached on this device.
What are the input limits?
Text sources are capped at 8,000 characters; image search accepts up to 20 images and a 500-character query. Models may truncate long text. First downloads can be hundreds of MB and need browser memory.