Encode conversations into a persistent knowledge graph with REST and MCP interfaces.
npx skills add https://github.com/ldclabs/anda-brain --skill anda-brain
Please help me install this Agent Skill. Skill: anda-brain Source: https://github.com/ldclabs/anda-brain/tree/main/anda_brain Command: npx skills add https://github.com/ldclabs/anda-brain --skill anda-brain
AI agents lose all memory after a conversation ends, and traditional storage—vector databases, Markdown files, and key-value stores—cannot maintain timeline, causality, or supersession relationships. Anda Brain provides a persistent Cognitive Nexus knowledge graph that compresses, evolves, and consolidates memories across sessions, ensuring agents remember the right information at the right time.
Use the anda-brain skill to persist user preferences and conversation history by sending messages to the Formation endpoint, then query the Recall endpoint with natural language questions to retrieve accurate, up-to-date memories.
High-intent search queries and answers about installing and using this skill.
Persistent long-term memory for AI agents is maintained by encoding conversations into a structured knowledge graph. This approach captures episodic, semantic, and cognitive memories, allowing agents to retain context across multiple sessions.
A vector database cannot maintain timeline, causality, or supersession relationships for AI agent memory. A cognitive nexus knowledge graph solves this by compressing, evolving, and consolidating memories to ensure accurate recall over time.
You query structured memories using natural language through a Recall endpoint. This interface synthesizes answers directly from the stored knowledge graph, returning up-to-date information grounded in episodic and semantic context.
Yes, REST and MCP interfaces support memory formation, recall, and maintenance operations with authentication. These interfaces enable multi-space isolation, making them suitable for business agents handling distinct user contexts.
A sleep cycle for memory maintenance is an automated process that consolidates, deduplicates, prunes, and decays stale knowledge in the graph. This ensures the AI agent retains relevant information while discarding outdated memory.
Update user preferences by sending new conversation data to the Formation endpoint. The knowledge graph applies supersession relationships, correctly remembering the current vegetarian preference while retaining the historical BBQ order context.
troykellyIntegrate episodic memory and knowledge graphs to maintain context across AI sessions.
javiarmestoAutomate adding generative AI nodes to Copilot Studio topics.
AradotsoCapture conversation turns and recall relevant context for MCP-hosted coding assistants.
davpatel605-beepStore and retrieve vector embeddings for semantic search and document retrieval.
