> ## Documentation Index > Fetch the complete documentation index at: https://docs.rippletide.com/llms.txt > Use this file to discover all available pages before exploring further. # Overview > Give your AI agents persistent memory with a context graph ## What is the Context Graph? The Context Graph is a persistent context graph that gives your AI agents memory across conversations. Without it, your agent forgets everything between sessions. With it, your agent can: * **Remember** facts, decisions, preferences, and context across sessions * **Recall** past information by searching the context graph * **Relate** entities to each other (e.g. "John works at Acme Corp") * **Invalidate** outdated information when things change Each agent gets its own **isolated context graph**, so memories from one project don't leak into another. ## When to use it * Your agent needs to recall past interactions or user preferences * You want structured knowledge (entities + relationships) instead of flat text * You need cross-session memory for coding assistants, support agents, or project tools * You want to plug memory into any MCP-compatible client (Cursor, Claude, VS Code) ## How it works Calls tools like `remember`, `recall`, `relate` automatically during conversations. The open transport protocol that connects AI assistants to the Context Graph. Think of MCP as the pipe. The Context Graph is what flows through it. Reads and writes to the agent's context graph. Stores entities, memories, and relations. Each agent gets its own isolated graph. You don't need to do anything. The assistant decides when to store or retrieve information based on context. ## Example Here is an interactive **live** environment where you can test the comparison. On the left, a standard agent without memory. On the right, the same agent with a real Context Graph via the playground proxy. For this Overview demo, the right side is preloaded with **code best practices**. Try asking: "How should I format a React component?"