As someone who is paralyzed themselves, I have had to use different forms of transcription throughout my education and work at different points and I have also met many people in the disabled community who have to use different forms of screen readers to be able to understand what is on the screen because they are either partly blind or suffer from intense dyslexia. Thinking about how AI has helped enable increased autonomy for many disabled people through adding natural language processing based ways of interacting with the world, I wanted to build a concept around the inverse direction too that both developers and disabled people could use to more easily experience and test where they needed to add alt text and then could easily generate it to do so while still being able to review it themselves. The result of this has been Alttext pro, an app based around generating and identifying alt text in both uploaded and observed webpages while also being able to control relevent accessibility features such as contrast,color,speed, and even two button interfaces directly from its settings. One of the major things I have learned through developing this process via medo is that the future will be one where increasingly developers and consumers alike can create apps for their particular needs more easily, but at the same time challenges like those around getting labeling functions to truely work will still be a key aspect that will need to be worked around even as AI has massively improved. That said, the increased ease of development of something like medo though it comes with the importance of needing to be aware of iteration will create a generation who think createing a personalized app for themselves is like doing so in scratch and in turn developers will feel more consciously aware of how they can contribute to accessibility too with their extra time.

Built With

  • gemini-vision-api
  • llms
  • medo
Share this project:

Updates