Concurrent writes are the right problem to be solving. Most agent memory implementations assume a single writer and fall apart the second you have parallel tool calls or multi-agent setups hitting the same store.
Curious how you're handling write conflicts at the semantic level though, not just the database level. Two agents writing contradictory facts concurrently is a different problem than two agents writing to the same row. Does the resolution happen at the DB layer or does the application need to handle it?
Interesting project! The single-writer limitation is a real pain point for multi-agent systems.
Worth mentioning ArcadeDB (https://arcadedb.com) — it's an open-source multi-model database (Apache 2.0) that supports concurrent writes natively, with graph (OpenCypher/Gremlin), document, key-value, and time-series models in one engine. No need to fork or maintain a separate project.
It also speaks the Neo4j Bolt protocol, so existing tooling works out of the box. Could be a good fit for agent memory use cases like this.