Aquin

floating AI assistant that co-works with you across all apps & sites, learns all the context so you never have to repeat yourself.

Comment

Inspiration

At 16, I took a 3 month break from school to build an AI browser, after building the MVP, I applied to VCs, most rejected me. But I realized that I could build something more out of it and ship it myself, where I picked the floating search engine inside the browser and built it system-wide which later become Aquin, floating AI assistant.

What it does

Aquin is a floating AI assistant that stays wherever you put it on your screen and works inside any app you're using. One input box handles voice input/output, file uploads of any type, browser tabs, generates inside any app or website you're typing in. The AI learns from every conversation, storing key insights to become more and more personalized. Aquin can also perform specific automation tasks like opening sites, apps, or manage your entire file system without taking over control. For privacy Everything can run locally with local LLMs with any Hugging Face model or by even connecting to Ollama.

How we built it

Built using Electron, TypeScript, React and Vite. Integrated Anthropic Claude for conversational AI, ElevenLabs for voice processing, Hugging Face for local models, and implemented MCP protocols for automation.

Accomplishments that we're proud of

Gained 15K+ YouTube followers and 150K+ views over youtube and X. Gave 100+ demos. Launched in beta. Gave interview with Grad Capital and ZFellows.

What's next for Aquin

  • Shipping Context Window Soon, Context Window are like a notepad plus browser that gives Aquin massive context.

  • Complete Google automation via MCPs (Gmail, Calendar, Meet, Drive) plus Notion and GitHub integration.

  • Exploring marketplace concept where anyone can create their own version of Aquin and publish it on a marketplace, think browser extensions but for your entire system.

  • jarvis from iron man like voice chat

Share this project:

Updates