Files, Context Windows, or Memory: Choosing Persistence for Production Agents with Memori Labs

Hosted by MongoDB Webinars

MongoDB and Memori Labs walk through the full agent memory lifecycle and compare three persistence options — larger context windows, files on disk, and a dedicated memory system on MongoDB — across concurrency, isolation, retrieval precision, governance, and cost. Includes a live build of Memori Labs writing trace and conversational memory into an Atlas cluster, benchmark results showing up to 36x token savings, and production concerns like supersession, per-tenant isolation, auditability, retention, and deletion.

When

Wed 2 Sep, 3:00 PM UTC

Format

Webinar

Speakers

Mikiko Bazeley (Staff Applied AI Engineer, MongoDB), Adam Struck (Co-Founder and CEO, Memori Labs)

For

engineers building production AI agents