The Problem: We save articles with good intentions, but keyword search fails when we don't remember the exact title, and "Read Later" apps just become another inbox we ignore.
The Solution: Memory Layer is a "set and forget" system. You click the extension, and the desktop app handles the rest.
How it works (The Tech):
Capture: Browser extension scrapes the readable content.
Processing: The Desktop app receives the content. It hits OpenAI's API to generate a summary and vector embeddings.
Storage: Data is stored in a local SQLite database on your machine.
Retrieval: When you search, we convert your query to a vector and perform a cosine similarity search against your local database to find the semantic match.
Privacy Note: While we use OpenAI for the processing (summarization/embedding), the actual database of your articles resides locally on your machine. We do not have a server database storing your articles.
Current Status:
Windows 10/11 only (Mac/Linux coming if there is interest).
Free during this early beta.
I am looking for brutally honest feedback. Does the semantic search actually find what you're looking for? Is the desktop app too heavy? Let me know.