SurrealDB launches Agent Memory system, scoring 77.2% on BEAM benchmark
tobiemh · reddit · 2026-10-15
SurrealDB founder Tobie announced Agent Memory, a system built on their multi-model database for developers building agents, context layers and knowledge graphs.
- Turns conversations and documents into structured entities, relationships and facts, with provenance and temporal history tracking
- Uses a single database supporting graphs, vectors and documents, avoiding stitching together multiple storage systems
- Benchmarked on BEAM with 77.2% accuracy; the team notes memory benchmark scores vary widely by evaluation protocol and chose not to chase the highest number
- Acknowledges alternatives like Mem0, Honcho and Cognee, but pursued a different approach
Related event: SurrealDB Launches Agent Memory, Scoring 77.2% on BEAM(4 posts)→
More from coding & agent
- Indie dev builds a MOBA in one week with Opus 5.5: 24 heroes, 3 battlegrounds — TAbrodi · 2026-10-15
- AI-built MOBA hits 13.1K plays and 11.3K players in its first week — TAbrodi · 2026-10-15
- Paige Bailey: Ship an Eval People Care About and You Own What Labs Hill-Climb On — DynamicWebPaige · 2026-10-15
- Voice-braindump to PRs: one dev's four-step workflow routing tasks across ten cloud agent projects — vinvan · 2026-10-15
- AI-rewritten libtiff closes gap to ~10x of C version, fueling hopes of retiring buggy C libs — jd_pressman · 2026-10-15
- DigiPet: open-source framework gives an LLM agent a cozy explorable desktop world — Nekileo · 2026-10-15