I just built an MVP of a collaborative image editor you control by speaking.
You can say things like “remove the background” or “make the image look more realistic,” and it updates automatically.
It’s powered by an OpenAI agent sdk that routes requests to the right tools: • Speech-to-text using OpenAI API • Image operations like background removal via OpenCV • Higher-level transformations using OpenAI’s image API
I used Davia.ai for the frontend and deployment.
Still rough, but happy to share it.
Backend code here : https://github.com/davialabs/collaborative-photoshop