Describe the world for blind and low-vision people
Be My Eyes built Be My AI on GPT-4 so blind and low-vision people can get instant image descriptions at no cost.
Built withChatGPT / OpenAI models
Try it withClaudeChatGPT / OpenAI modelsGeminiGrok
Deployed, Be My Eyes
Why it matters
Reading a label, a menu or a form is a daily barrier for blind people. An instant describer that costs nothing adds to the existing volunteer network.
How it works
Users send a photo and the assistant describes it and answers follow-up questions; a human volunteer is still available (Be My Eyes). TechCrunch reported in March 2023 that the 'Virtual Volunteer' was the first app to integrate GPT-4's image capability. Be My Eyes said at the time that it had 6.3 million volunteers.
Who it's for
Blind and low-vision people, families and accessibility teams.
The prompt
You are a visual assistant for a blind user. For each photo: 1. Start with a one-sentence summary of what it is. 2. Read any visible text exactly, in reading order. Say clearly if text is cut off or blurry. 3. Describe the layout, with positions (top left, centre and so on) and any safety-relevant details such as steps, traffic or hot surfaces. 4. If you are unsure about anything (expiry dates, medication names, amounts of money), say so and suggest the user confirm with a person. For medicines, read the label text exactly but give no dosing or medical advice; suggest checking with a pharmacist. Keep answers to two or three sentences, then ask if I want more detail.
Open inClaude (opens in a new tab)ChatGPT (opens in a new tab)Gemini (opens in a new tab)Grok (opens in a new tab)
Sources
1. Be My Eyes: Be My AI powered by GPT-4; no cost to users. Introducing Be My AI (opens in a new tab)
2. TechCrunch: Virtual Volunteer launch on GPT-4. GPT-4's first app is a virtual volunteer for the visually impaired (opens in a new tab)