Badge

Badge: a wearable conference badge that remembers who you meet. The problem At conferences I meet lots of people and need to remember them. The solution Badge is a voice first badge you wear on a lanyard. After a conversation, I tap record and talk. Badge transcribes the audio, creates a contact with name, role, and company, writes a summary with key highlights, and lists next steps: an email to draft, a number to call, or a meeting to schedule. Everything is saved so I can follow up while the conversation is still fresh. Nothing sends without my review. How it works A Raspberry Pi 5 with a 3.5 in screen and a USB microphone runs the web app, in a standard case on a lanyard I knitted myself. The app is built with React, Vite, a Node.js server, and Supabase, and it’s deployed on Vercel. Use of Meta AI tooling Meta Muse Voice Transcribe turns the recording into a transcript. Meta Muse Spark extracts the contact details, writes the summary and highlights, and finds the next steps. Privacy Audio is deleted after transcription. Each user can only access their own data, API keys stay on the server, and there is no face recognition. Conversations should only be recorded with permission. Who it helps Anyone whose work runs on relationships, such as founders, salespeople, recruiters, and researchers. It could also help people who struggle to remember names and faces. What’s next Grouping notes under one contact, email and phone fields, Gmail and Calendar links, and an “ask your network” search across everyone I’ve met.