Memoir
Memoir is an AI-powered collaborative meeting assistant that listens to spoken prompts and updates a shared canvas in real time. It lets users privately pull up files, draw diagrams, or retrieve call details by voice, with content staying hidden until explicitly shared. The system works during the conversation, refining its output continuously without requiring typing.
- AI Agents
- Enterprise Software
- Artificial Intelligence
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
Traditional virtual meetings rely on separate tools for note-taking, file sharing, and diagramming, forcing participants to switch contexts and manually capture decisions. This fragmentation slows down discussions, loses important details, and makes it difficult for everyone to stay aligned in real time.
Solution
Memoir provides a call with a shared canvas and an AI agent that listens to everyone who lets it. Users speak naturally to prompt actions: saying "let me pull up the deck" makes the file appear privately on their screen, and pressing Return shares it with the room. Saying "let's draw it out" generates boxes and labels on the shared canvas as the user talks, and anyone can move them. The agent also answers questions from the call itself, such as "what did we say about pricing?" with a short, cited answer. Memoir works in real time, starting with quick drafts and refining them with a slower, more careful pass, ensuring nothing on the call waits for the final version.
Target Audience
Primary users are founders and their teams who need a collaborative, real-time meeting assistant to improve discussion flow and capture decisions without manual note-taking.
Features
- Voice-driven interaction: users speak to pull up files, draw diagrams, or ask questions, with no typing required
- Private-then-share workflow: content appears on the user's screen alone until they press Return to show it to the room
- Real-time canvas: boxes, labels, and diagrams appear on a shared canvas as users talk, and anyone can move them
- Just-in-time processing: first marks land about a second after words, with a slower, more careful pass refining the output when ready
- Screen capture: users can say "grab my screen" to place a screenshot of their real app on the canvas, with changes drawn on top
- Privacy controls: never speaks uninvited, never shows private material until shared, never hears opted-out participants, and never trains on calls