Conversation assist app created with ai and decentralized storage.
At large conferences like Devcon, it's common to meet new people, take a selfie or PFP (profile picture), and post it on Telegram. However, it’s challenging to keep track of the conversations you had or even remember the face of someone who might have shared a great idea. Over time, these connections often fade away.
Introducing Convo, an AI assistant designed to help you track your conversations, summarize key points, and provide valuable insights.
As a proof of meeting, we securely store the PFP—either privately or publicly—on decentralized storage. The choice is yours.
To ensure privacy and security, the entire local system is verified using Trusted Execution Environment (TEE) technology.
This app was built in just 12 hours. After two of my teammates dropped out at the last minute of the ETHGlobal hackathon, I decided to carry forward our original ideas and pivot them into this concept.
The app extensively uses Filframe. It is deployed on Filecoin using Forge and Filframe’s native script for deployment. I also deployed it on Morphe and utilized Phala Network for TEE (Trusted Execution Environment) on my local server.
The frontend is built with Next.js.
For the hardware device integration, I used Omi. I implemented a webhook receiver to process incoming data and leveraged OpenAI for advanced analysis.