Observable memory and caching for your AI agents
BetterDB is the open-core memory and caching layer for AI agents, running on any Valkey or Redis you own.Memory gives your agents durable, scoped recall: remember and recall with ranking by similarity, recency, and importance, so an agent carries context across sessions instead of starting cold. Caching answers repeat work fast: a semantic cache catches paraphrased hits, amulti-tier exact-match cache serves LLM responses, tool results, and session state in under a millisecond. The difference from a hosted black box is that you see every decision on a dashboard and can change how it behaves while the app is running. Thresholds,TTLs, invalidation, all observable and adjustable live. The cache can even tune itself: an MCP agent reads live state and proposes a change, a human approves it, and the running app picks it up in seconds with no redeploy. Run it self-hosted on your own Valkey, or fully managed. Either way, no vendor lock-in. MIT core, Type Script and Python. Live demo at chat.betterdb.com.
We send new OSS products every week in a new newsletter. No Spam.
Approximately, we add new tools within three months.
We will publish it with a no-follow link.
However, you can publish your tool immediately and get a forever do-follow link.
See you soon.