The AI Agent Memory Landscape
summary
Argues that context graphs plus three concrete memory types — short-term, long-term and reasoning memory — are the practical difference between demo agents and production systems, and ships a Neo4j agent memory Python package that integrates with SpaCy, local GLER models, Anthropic Sonnet, OpenAI embeddings and graph data science. Demonstrates a working create-context-graph scaffold (Next.js frontend, seeded demo data, connectors for Cloud Code, GitHub, Linear, Google Workspace, Aura) and shows heuristics that extract decision traces; TCK tests include six cross-language agents and a 16-project demo.
AI-written summary · about
chapters (9)
- 0:00Introduction and setup
- 1:12Context graphs and agent memory
- 3:19Demo financial services context graph
- 9:57Create Context Graph project
- 15:20Demo science research graph
- 18:51Demo cloud code connector
- 20:17Multi-agent memory and coordination
- 23:48Agent Memory TCK and future
- 25:15Resources and closing
related talks
51mMemory Masterclass: Make Your AI Agents Remember What They Do!
23mConnecting the Dots With Context Graphs
15mBuilding Evolving AI Agents Via Dynamic Memory Representations Using Temporal Knowledge Graphs
26mStore Your AI Agent's Memory and Context in a Knowledge Graph
56mConnecting the Dots with Context Graphs
20mWhy your agents need decision traces, not just documents
46mDesigning Semantic Memory for Multi-Agent Systems with Python
18mConnecting the Dots with Context Graphs