claude-devtools is a debugging tool for Claude Code that provides visual inspection of AI agent sessions, tool calls, token usage, and context window state.
The tool solves the problem of Claude Code's opaque output by reading session logs directly from the Claude Code installation on your machine and presenting detailed information in a visual interface. When Claude Code replaced granular output with summary text, developers lost visibility into what the AI agent actually did—which files it read, what it searched for, how many tokens it consumed, and how it structured its reasoning. claude-devtools restores this visibility by parsing the underlying session data and displaying session transcripts, tool call details, token usage metrics, and subagent activity in an interactive UI.
The tool suits developers who need to debug Claude Code's behavior, understand token consumption patterns, or audit what actions the AI agent performed during a coding session. It works for anyone running Claude Code locally and wanting deeper observability into agent decision-making and execution. The tool runs as a desktop application on macOS, Linux, and Windows, and can also be deployed via Docker for standalone use.
Development activity shows consistent engagement with the problem space. The project maintains active responsiveness to issues and pull requests, indicating the maintainer addresses user-reported problems and accepts community contributions. The codebase demonstrates ongoing refinement of the debugging interface and session parsing logic based on feedback. The tool has accumulated community recognition within the Claude Code ecosystem, suggesting sustained interest and validation of the core value proposition.
