What is Sno?
Sno is the second brain for your AI agents. Just as you use tools like Notion or Obsidian to organize your thoughts, your AI agents (Claude, ChatGPT, Cursor) need a persistent, structured way to remember context across sessions. Sno provides this memory layer.
Why Sno?
🧠 Persistent Context
LLMs are stateless by default. Sno gives them long-term memory. When you tell Claude about your project architecture today, it remembers it next week.🕸️ The Memory Graph
Sno doesn’t just dump text into a database. It organizes information into a 3D Knowledge Graph, connecting related concepts (e.g., “Project X” is related to “React” and “Deadline Y”). This mimics how human brains work.🔌 Universal Protocol (MCP)
Sno is built on the Model Context Protocol (MCP). This means it works out-of-the-box with any MCP-compliant tool, including:- Claude Desktop App
- Cursor / Windsurf
- Zed Editor
- Custom AI Agents
How it Works
- Capture: You interact with your agent naturally.
- Process: The REM Agent runs in the background, analyzing your conversation to extract facts, decisions, and relationships.
- Store: These insights are stored in your private Memory Graph.
- Recall: In future conversations, Sno proactively injects relevant context before the agent even answers.