> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://help.getzep.com/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://help.getzep.com/_mcp/server. # What is context engineering? > Context engineering assembles the right information around an LLM for reliable agents. Zep does it with temporal knowledge graphs and Smart Context Assembly. Context engineering assembles information, instructions, and tools for a large language model (LLM). Prompt engineering focuses on instructions, while context engineering also supplies data and tools. LLMs are stateless outside their context window. An application must provide the background knowledge, user history, business data, and tool output for each request. Zep constructs a [Context Graph](/graph-overview) from [business data](/adding-business-data) and [chat messages](/adding-messages). A Context Graph is a temporal knowledge graph for an object, system, or user. The graph contains entities, relationships, and facts. When new data contradicts a fact, Zep can invalidate the prior fact while preserving its history. Use [retrieval methods](/assembling-context) to supply selected graph data to an agent.