kkkirito-123/zlagent

ZLAgent is a personal AI assistant that lives inside messaging apps, specifically WeChat, WeCom, the enterprise version of WeChat, or any webhook-compatible chat platform. Instead of opening a separate app or website to talk to an AI, you chat with it like a contact right inside the IM platform you already use. The assistant is designed to grow smarter over time through three kinds of stored knowledge: long-term memory, your preferences and habits, stored in PostgreSQL, an answer and fact cache for repeated questions it has already worked out, managed in a wiki subsystem, and a knowledge graph of structured relationships extracted from conversations and documents, stored in Neo4j. Together these let it remember you across sessions, skip recomputing answers it already knows, and follow connections between related facts. It also has a skill and tool system. Fifteen built-in tools cover things like memory management, scheduling tasks, web search, reading URLs, running code, and sending messages. You can extend it further by installing MCP tools, a standard for connecting AI agents to external software, through a natural-language request in the chat, with no configuration files to edit manually. Tools have three permission levels: safe ones run automatically, confirm-level ones pause and ask you before running, and deny-level ones are blocked entirely. The whole stack runs via Docker with a single command, spinning up the API server, PostgreSQL, Neo4j, and Redis together. The primary language is Python, using FastAPI for the web layer, and the project is released under the MIT License. The full README is longer than what was shown.