A shared brain for your coding agents

Maximum context for your coding agents - shared across IDEs, projects and teams

Compatible with

Why do your coding agents still feel inefficient?

Agents always start from scratch

Hundreds of hours lost re-prompting, re-specifying rules and context for each new session.

Knowledge stuck in one project or IDE

Lose all your coding context when switching projects, tools, or environments.

Agents don't learn from each other

Valuable learnings get lost; your team's agents keep making the same mistakes, again and again.

Now, with Cipher all of that changes

Auto-Generated, Auto-Retrieved Memory

Your AI agent's memory is generated and retrieved—scaling with your codebase.

Bug fixes, reasoning steps, and best practices—all instantly stored and retrieved, while you code.

Cross-IDE Coding With Shared Memory

One memory layer, everywhere you code.

Cipher connects Cursor, ClaudeCode, VS Code, and more. Your project memory and context travel with you.

Team-Shared Memory

Build a single source of truth for your team.

Share memory so every agent and teammate learns from everyone's best practices, instantly.

Install and Run In Two Minutes

Install Cipher with zero configuration.

Connect to any IDE through MCP integration. Get started in under 2 minutes.

Frequently Asked Questions

What is Cipher?

Cipher, built by Byterover (@byterover.dev) is an open-source memory layer specifically designed for coding agents compatible with multiple AI IDEs through MCP.

What IDEs and coding agents does Cipher support?

Cipher is compatible with all major AI IDEs including Cursor, Claude Desktop, Claude Code, Windsurf, Gemini CLI, AWS Kiro, VS Code, Roo Code, and many more. Through our MCP integration, Cipher works with virtually any coding agent or IDE.

How can I contribute to Cipher?

To contribute to Cipher, please read our CONTRIBUTING.md and the issue templates in .github/ISSUE_TEMPLATE in our Github repo and follow the guidelines when filing issues or making changes, and open a pull request against main.

Does Cipher support self-hosted LLM?

Yes, Cipher is now compatible with Ollama, LM Studio and OpenRouter, allowing user can develop with opensource models.

Still have questions?

Our community and support team are here to help you get the most out of Cipher.

Thousands of dev teams are transforming how they code with AI with the memory layer.