LangSmith launches SmithDB, a distributed database designed specifically for agent observability

AIMPACT News, May 14 (UTC+8), at the Interrupt! conference, LangChain announced the launch of SmithDB, a distributed database built on Rust, Apache DataFusion, and Vortex, focused on handling large-scale, complex agent tracking data (including multimodal content and long-span scenarios). Its architecture consists of object storage, a small Postgres metadata store, and stateless services, requiring no management of local disks, supporting elastic scaling and self-hosted deployment. In terms of performance, the tracking tree load P50 latency is 92ms, a single run load is 71ms, run filtering is 82ms, full-text search is 400ms, thread filtering is 131ms, with core LangSmith experience speed improvements of up to 12 times. Currently, SmithDB is fully serving production traffic, covering 100% of US cloud ingestion and tracking UI queries, with product integration (rule execution, batch export, experimental features) nearing completion. (Source: InFoQ)

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned