Building a personal AI that respects your privacy

The promise of personal AI - an assistant that knows your preferences, projects, and patterns - requires memory. And memory requires data. The question is: where does that data live?

The cloud memory trap

Most AI products store your conversations, files, and context on their servers. They argue it's necessary for features like memory, search, and personalization. But it's not. Every one of those features can run locally - and we've built them that way.

How Gotchi's memory works

• Conversations are stored in an encrypted local database. Only you can access them.
• The knowledge base indexes your files into a local vector store for semantic search.
• Personal memory extracts preferences and patterns from conversations and stores them on-device.
• Connected apps (Gmail, Slack, etc.) sync to a local index - never through an intermediary.

The result is an AI that genuinely knows you - your writing style, your projects, your preferences - without a single byte leaving your hardware. It's personal AI that's actually personal.

Privacy isn't a feature we bolted on. It's a constraint we designed around. Every architectural decision starts with: can this run entirely on the user's machine?