architecture
Per-User Memory for AI Products: Multi-Tenant Patterns
How to give every user isolated memory in a multi-tenant AI product: hard boundaries vs soft partitions, testing for cross-tenant leakage, GDPR deletion, and scaling.
One Bank or Many? A Field Guide to Structuring Agent Memory
Every Hindsight integration tells you to set a bank_id, but not how to decide what a bank should be. A bank is a recall boundary: how to scope agent memory, and when to reach for tags instead.
The Consolidation Problem in Agent Memory
Agents that remember everything remember nothing useful. A four-lever framework — importance, merge, decay, eviction — for consolidating agent memory.
The Case Against External Vector DBs for Agent Memory
Most 'add memory to your agent' tutorials start with installing a vector database. They shouldn't. Why vector DBs are the wrong default for agent memory.
How Hindsight Scales
A design analysis of how Hindsight's memory operations scale with data volume — what costs grow, what stays bounded, and why.