Memory Time Machine Talk to Your Past.

We document our lives more than ever before — but we rarely understand them.

Every day, people generate massive amounts of personal information: journal entries, voice notes, screenshots, photos, unfinished goals, scattered thoughts, saved articles, and emotional moments.

But traditional note-taking apps only store information. They don’t understand it.

We wanted to explore a deeper question:

What if your memories could become intelligent?

That idea became Memory Time Machine — an AI-powered personal memory system designed to transform fragmented life experiences into a searchable, emotionally aware archive.

Users can upload memories through text, voice memos, images, files, and webpages. The system then uses large language models, OCR, speech-to-text, vector embeddings, and conversational AI to extract emotional patterns, summarize experiences, identify commitments, and build contextual understanding over time.

But the most powerful part isn’t storage.

It’s reflection.

With Memory Time Machine, users can literally talk to their past self through natural language conversations:

“What habits are holding me back?” “What promises do I keep breaking?” “When was I happiest?” “What emotional patterns do you notice?”

The platform retrieves relevant memories, synthesizes insights, and responds conversationally with contextual citations.

We also built specialized AI insight systems, including:

a Broken Promises Detector that identifies recurring commitments users fail to complete, and a Future Self Simulator that predicts where current habits and emotional patterns may lead over time.

Technically, the project combines:

React + TypeScript, Supabase, pgvector semantic search, Gemini 2.5 Flash, Whisper speech recognition, OCR pipelines, and serverless AI workflows.

We used MeDo to accelerate ideation, rapidly iterate on workflows, experiment with conversational product design, and streamline full-stack development.

Memory Time Machine is more than a productivity tool.

It’s an experiment in helping people better understand themselves through AI.

Built With

• css
• postgresql
• react
• supabase
• typescript
• vite