Cipher Logo
Cipher is an opensource memory layer specifically designed for coding agents. Compatible with Cursor, Windsurf, Claude Desktop, Claude Code, Gemini CLI, VS Code, and Roo Code through MCP, and coding agents, such as Kimi K2. Key Features:
  • Connect with your favorite IDEs through MCP
  • Auto-generated memories that scale with your codebase
  • Dual Memory Layer that captures System 1 (Programming Concepts & Business Logic & Past Interaction) and System 2 (Reasoning Steps of the model when generating code)
  • Install on your IDE with zero configuration needed
  • Switch seamlessly between IDEs without losing memory
  • Shared memory workspace across the team in real time

​
How Cipher Works

​
IDE Integration through MCP

Cipher works with any coding tool that supports the Model Context Protocol. This means you can use it with Cursor, Claude Desktop, Claude Code, VS Code, and more without switching tools or losing your workflow.

​
Auto-Generated Memories

Cipher automatically creates and organizes memories from your coding sessions without any manual input. As your codebase grows, the memory system scales with it, continuously learning from your development patterns and storing relevant knowledge that becomes more valuable over time.

​
Dual Memory System

Cipher captures two types of knowledge to make your AI assistant smarter:
  • System 1 Memory: Remembers programming concepts, business logic, and your past interactions. It learns what your code does and how you prefer to work.
  • System 2 Memory: Captures the reasoning steps the AI uses when generating code. This helps the AI learn from its own problem-solving process to give better suggestions over time.

​
Zero Configuration Setup

Just add Cipher to your development environment and it starts working immediately. No complex configuration files or setup processes required - it integrates seamlessly with your existing workflow.

​
Cross-IDE Memory Persistence

Your AI assistant remembers everything whether you’re in Cursor, Claude Desktop, or any other supported tool. Start a conversation in one IDE and continue it in another without losing context.

​
Team Collaboration

Team members can benefit from each other’s interactions with the AI, building collective knowledge about your codebase and development practices. This creates a shared understanding that helps everyone work more efficiently.

​
Technical Architecture

Cipher Architecture Diagram Cipher’s architecture is built around a dual memory system that captures both explicit knowledge and reasoning patterns: System 1 Memory (Explicit Knowledge)
  • Programming Concepts: Stores information about your codebase architecture, patterns, and business logic
  • Past Interactions: Remembers previous conversations and solutions to build context over time
  • Vector Search: Uses embeddings to find relevant knowledge quickly when you need it
  • Knowledge Graphs: Maps relationships between code concepts, files, and development decisions
System 2 Memory (Reasoning Patterns)
  • Reasoning Steps: Captures how the AI thinks through problems and generates solutions
  • Problem-Solving Patterns: Learns from successful approaches to apply them to similar challenges
  • Code Generation Logic: Understands the thought process behind code suggestions
  • Continuous Learning: Improves suggestions by analyzing what works and what doesn’t
Integration Layer
  • MCP Protocol: Connects to coding assistants and development tools seamlessly
  • Multi-Provider Support: Works with OpenAI, Anthropic, OpenRouter, and local Ollama models
  • Flexible Deployment: CLI, API server, or MCP server modes for different use cases

​
Next Steps

  1. Quick Start - Get cipher running in under 5 minutes
  2. Connect Your Tools - Integrate with Cursor, Claude Desktop, Claude Code, and more
  3. Configuration - Customize cipher for your development workflow