Ourmem FAQs

Name: Ourmem
Author: ourmem

Question 1

What core problems does Ourmem solve for AI agents?

Accepted Answer

Ourmem eliminates AI agent 'amnesia' by providing persistent memory, breaks knowledge 'silos' between agents, enables cross-device continuity, facilitates team-wide sharing, and upgrades 'dumb recall' to deep semantic understanding for collective intelligence.

Question 2

How does Ourmem ensure accurate and relevant memory retrieval?

Accepted Answer

It features an advanced 11-stage hybrid retrieval pipeline combining vector search, BM25 keyword matching, RRF fusion, cross-encoder reranking, decay model boosting, and MMR diversity to achieve deep semantic understanding and precise recall.

Question 3

How does Ourmem manage the lifecycle of memories?

Accepted Answer

It employs a Weibull decay model to automatically manage memory lifecycles, ensuring core memories persist while peripheral ones gracefully fade. Memories are promoted or demoted across Core, Working, and Peripheral tiers based on access and importance.

Question 4

What types of data can Ourmem ingest and process?

Accepted Answer

Ourmem offers multi-modal ingestion capabilities, supporting various data types including PDF documents, image OCR, video transcription, and detailed code AST chunking, allowing agents to learn from diverse sources.

Question 5

How does Ourmem support knowledge sharing across different boundaries?

Accepted Answer

Ourmem utilizes a three-tier Space architecture (Personal, Team, Organization) allowing knowledge to flow across agents and teams. It includes provenance tracking and quality-gated auto-sharing rules to ensure relevant insights are shared.

Ourmem

Ourmem

主要功能

使用案例

主要功能

使用案例