Memori is an LLM-agnostic memory layer that captures, classifies, and recalls structured state from agent conversations. It automatically extracts facts, preferences, and summaries, provides targeted recall with semantic search, and delivers explainable results. Supports role-based access control, data retention policies, and audit trails. Benchmark accuracy of 81.95% with 95% token reduction.
Freemium
19
How to use Memori?
Integrate Memori by adding the SDK with one line of code. It automatically captures each chat turn and classifies it into facts, preferences, rules, and summaries. Use targeted recall to pull relevant context across conversations. For production, choose a plan that fits your usage. Self-host with your own database or use Memori Cloud for hosted storage and search.
Memori 's Core Features
Automatically classifies chat turns into facts, preferences, and summaries
Targeted recall across conversations and documents
Selective semantic search with token cost optimization
Explainable results with entity, time, and source lineage
Role-based access control and data retention policies
Benchmarked 81.95% accuracy on LoCoMo with 95% token reduction