Skip to content

A voice-powered journaling app built with a RAG retrieval system to gain insights from your journal entries

Notifications You must be signed in to change notification settings

andrew-fenton/memora

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

65 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

memora

memora_thumbnail

memora is a journaling app that knows you better than yourself. It works as a voice diary, allowing you to record spoken diary entries. memora summarizes and analyzes these entries to understand your personality, traits, and values. With this info, you can ask memora any question such as "do I overwork myself?" or "Am I improving at collaboration?" and it can give you thorough answers based on your entries, allowing you to more deeply understand yourself.

the inspiration.

My phone contains over 2GB of voice memos, featuring countless moments of meaningful introspection, happy moments, and forgotten stories, but none of them are ever replayed once they've been recorded. In the search of figuring out a more efficient and helpful way for people to record and store voice diaries, the team realized that data could be used rather than just stored. From that thought, they came up with memora, which summarizes recordings to make them easy to look back at, while also helping people understand themselves on a deeper level.

the technical details.

  • Voice-to-Text Transcription: Built a transcription layer using Google's Speech API to produce accurate speech-to-text conversion while retaining user intent.
  • Dual Database Architecture: Used MongoDB for raw journal entries to support flexible, unstructured data, and ChromaDB for efficient vector storage and retrieval in our RAG setup.
  • Vector Search with Cohere: Integrated Cohere’s command-xlarge-nightly model for embedding entries into high-dimensional vectors, enabling semantic searches to pull contextually relevant results beyond exact keyword matching.

what's next for memora.

lots :)

About

A voice-powered journaling app built with a RAG retrieval system to gain insights from your journal entries

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published