claudeye is an easy-to-use application that helps you monitor and review sessions from Claude Code & Agents SDK. If you use Claude agents or code, claudeye lets you replay sessions, run your own evaluations, and debug agent traces. It helps you understand what your agents do and how they behave to make better decisions and improve your work.
You don’t need special skills or programming knowledge to start using claudeye. It works as a tool to watch, analyze, and troubleshoot your Claude-related work simply and clearly.
Before installing claudeye, make sure your computer meets these basic requirements:
- Operating System: Windows 10 or later, macOS 10.14 or later, or any recent Linux distribution
- Processor: At least 1.5 GHz dual-core processor
- Memory: Minimum 4 GB RAM
- Disk Space: At least 200 MB free space
- Internet Connection: Required for downloads and updates
If your computer meets these, you should have no problem running claudeye smoothly.
To get started with claudeye, visit this page to download the version that fits your system:
- Click the link above to open the claudeye releases page.
- Look for the latest release version. You’ll see files for different operating systems.
- Choose the file that matches your system:
- For Windows, look for a file ending with
.exe - For macOS, look for a
.dmgor.zipfile - For Linux, look for
.AppImageor.tar.gz
- For Windows, look for a file ending with
- Click the file to download it.
- Once downloaded, open the file and follow the on-screen instructions to install claudeye.
- After installation, find the claudeye app icon on your desktop or start menu and open it.
If you have any trouble during download or installation, try running the installer as an administrator or check your internet connection.
After installing, launch claudeye and you’ll see a clean, simple interface designed to help you explore your Claude SDK sessions.
-
Session Replay
Watch recorded Claude Code & Agent sessions step by step. This helps you see what actions agents took and when. -
Custom Evaluations
Run your own tests to check how agents perform using your criteria. -
Debug Agent Traces
Use these logs to find problems or unexpected behavior in your agents.
- Open the app.
- Use the menu to select a session you want to review.
- Click play to see the agent’s activity in real-time.
- If you want to create a custom evaluation, select "New Eval" and follow the prompts.
- Explore the agent traces within the interface to troubleshoot or improve agents.
The interface uses clear labels and buttons to guide you through these tasks.
- Always save your sessions after recording them to avoid missing important data.
- When creating custom evaluations, keep your criteria simple and focused for clearer results.
- Use the debug traces to understand errors or unexpected agent responses—they show detailed info step by step.
- Check the claudeye updates regularly on the releases page to get new features or fixes.
- If you use command line tools with Claude SDK, claudeye can complement them by giving a visual understanding of what happens.
If you want to deepen your knowledge and get tips on how to use claudeye most effectively, explore these topics:
- Basics of Claude Code and Claude Agents SDK
- How agent sessions work and what you can track
- Writing meaningful custom evaluations
- Understanding debug and trace logs
- Best practices for troubleshooting agents in different scenarios
If you have questions, find bugs, or want to suggest improvements:
- Check existing issues or open a new one on the GitHub repository
- Visit the claudeye Discussions section for ideas and answers
- Review documentation linked in the repository for additional tools and help
Your feedback helps improve claudeye for everyone.
- Name: claudeye
- Description: Watchtower for Claude Code & Agents SDK - replay sessions, run custom evals, debug agent traces. Uncover, Understand, and Utilize.
- Topics: ai-agents, claude, claude-agents-sdk, claude-code, cli, eval, eval-framework, npm, observability, reliability, reliability-engineering