$ loading_
Provide layered long-term memory for AI agents across markdown, vector, and graph stores.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "sandbook" yet — see the docs or source repo.
Explain how to connect sandbook as an MCP memory service to a Claude agent, and design a memory structure using markdown notes, vector retrieval, and graph relationships.
A setup guide and a recommended layered memory architecture for the agent.
Use sandbook to design a long-term memory plan for a software project, storing requirements, decision logs, technical docs, and entity relationships, and explain how to retrieve and update them later.
A project knowledge base plan with storage layers, retrieval methods, and maintenance workflow.
Design a multi-turn conversation memory strategy with sandbook: store short-term information in markdown, embed key facts into vectors, write people and event relations into a graph, and provide usage rules.
A clear set of memory write and read rules to help the agent maintain conversational continuity.
Provide AI agents a local isolated Linux sandbox for fast, safe commands.
Provide AI agents a local sandbox to run code and tasks safely.
Run commands, manage long jobs, and transfer files in AI sandboxes.
Provide persistent graph-memory storage and retrieval for LLM agents via MCP.
Add per-repo persistent memory with versioned retrieval for Claude Code.
Give AI assistants persistent, searchable memory through an Obsidian-like knowledge system.