Databaset
Persistent memory for AI applications
I rebuilt the same AI memory system four different times.
Every AI app I worked on needed the same thing, remember what a user said, recall it later when relevant. Every time, same story. Set up a vector database. Write chunking logic. Wire up embeddings. Handle the edge cases where old info contradicts new info. Two weeks gone before I even touched the actual product.
Fourth time, I stopped and asked why I was rebuilding this from scratch again.
So I built Databaset. Two lines of code now. Store a memory, recall it later, done. No vector DB to configure, no chunking script to write, no contradiction logic to handle manually.
Free tier is live if anyone wants to try it. Mostly looking for people who've hit this same wall to tell me what's still missing.
1 Comment
I like the origin story. Rebuilding the same infrastructure four times is usually a sign it should be its own product. The best developer tools often come from someone getting tired of solving the exact same problem over and over again.