Key Facts
- ✓ The tool is a single Go binary that syncs session history to SQLite.
- ✓ It supports TUI, CLI, and MCP interfaces.
- ✓ Default mode is a TUI with a session browser and full-text search.
- ✓ Users can resume sessions or export them to Markdown.
- ✓ The MCP server allows Claude to search previous sessions for context.
Quick Summary
A new utility named ccrider has been introduced to manage Claude Code sessions effectively. Developed by Neil Berkman, this tool addresses the need for persistent conversation history by syncing session data to a SQLite database every time it is used.
The tool is designed as a single Go binary, ensuring lightweight and efficient operation. It supports three primary interfaces: a Terminal User Interface (TUI), a Command Line Interface (CLI), and a Model Context Protocol (MCP) server. The TUI serves as the default mode, offering a session browser and full-text search functionality. Users can navigate their history, search within specific sessions, resume conversations, or export data to Markdown format. The MCP server integration allows Claude to access this historical data, enabling it to retrieve context from previous sessions automatically.
Core Functionality and Interfaces
The ccrider tool is built to solve the problem of context retention in AI coding sessions. By maintaining a complete history, it prevents the loss of valuable information during long development cycles. The architecture relies on a SQLite backend, which is updated with every interaction to ensure data integrity.
The tool offers a versatile set of interfaces to suit different workflows:
- Terminal User Interface (TUI): This is the default mode, featuring a visual session browser and robust full-text search capabilities. It allows users to visually navigate their history and inspect session details.
- Command Line Interface (CLI): Provides direct access to tool functions via terminal commands, ideal for scripting and automation.
- Model Context Protocol (MCP) Server: This interface connects directly to Claude, providing tools that allow the AI to search through previous sessions for context or pull data from prior interactions.
Once a specific session is identified, users can perform several actions: browse the conversation, search for specific text within the session, resume the session directly, or export the entire conversation to a Markdown file for documentation or sharing purposes.
"I use this constantly. I've seen elaborate continuity systems to give Claude Code access to history but this simple approach has been very effective."
— Neil Berkman, Developer
Installation and Setup
Installation of the tool varies depending on the operating system. The developer has provided specific instructions for macOS and Linux users to ensure a smooth setup process.
For macOS users, the installation is handled via the Homebrew package manager. The command to install the tool is:
brew install neilberkman/tap/ccrider
For Linux and other operating systems, the installation requires building the tool from the source code available on GitHub. The steps involve cloning the repository and compiling the Go binary:
git clone https://github.com/neilberkman/ccrider && cd ccrider && go build
To utilize the MCP server functionality, a specific configuration command must be executed within the Claude environment. This command registers the tool with the user's scope:
claude mcp add --scope user ccrider $(which ccrider) serve-mcp
Developer Insights and Availability
The tool was created by Neil Berkman and released on GitHub. It was also shared via the Y Combinator news platform. The developer noted that while there are elaborate systems designed to give Claude Code access to history, this specific approach focuses on simplicity and effectiveness.
According to the developer, the tool is used "constantly" and has proven to be very effective despite its simple design. The source code and repository are publicly available, allowing users to inspect the code or contribute to its development. The project is hosted under the username neilberkman on GitHub.
