Hi Product Hunt! I’m Nicolas, a puppeteer and automata maker. I started building Bubblai because I was tired of copying prompts between conversations and combining separate tools whenever I wanted to shape an AI character properly. I wanted to see what makes up an agent’s mind. In Bubblai, the conversational agent sits at the centre. Around it, you connect visual bubbles for its role, personality, instructions, memory, documents, voice, web access, image and video generation, tools and actions. You can connect a module, test what changes, and remove it whenever it no longer belongs. The idea is to create a specialised assistant, a carefully crafted companion, a role-playing character or an agent capable of acting — without hiding everything inside one enormous prompt. Bubblai can also communicate with connected objects and robots, but that is only one possible extension of the system. I’d genuinely like to know whether this visual way of composing an AI feels clear when you first try it, and what still feels confusing or missing.

Bubblai

Compose characters and AI agents with visual modules

Compose characters and AI agents with visual modules