Where are Claude Code transcripts located?

Transcripts are automatically stored by Claude Code in the .claude/logs/ directory of your project, typically organized into subdirectories by date.

Does it include token usage statistics?

Yes, by using the --include-usage flag, you can append token usage and cost statistics to your condensed summary.

Does this skill support batch processing?

Absolutely. You can point the transcript-condenser at a directory of logs to process multiple sessions at once, saving the results to a specified output directory.

Can I export logs to a specific file format?

Yes, the skill supports both Markdown (best for human reading) and JSON (best for programmatic analysis) formats using the --format flag.

How can I see only the tools Claude used during a session?

You can use the --only-tools flag to filter the output, displaying only the assistant messages that involved tool execution and their results.

Transcript Condenser

Name: Transcript Condenser
Author: OrasesLabs

byOrasesLabs

Analytics & Monitoring

Summarizes verbose Claude Code session transcripts by filtering system noise and metadata into human-readable logs.

About

The Transcript Condenser skill transforms cluttered Claude Code logs into clear, actionable summaries, allowing developers to quickly review tool usage, subagent interactions, and session outcomes. By stripping away system artifacts and providing multiple verbosity levels—from minimal overviews to detailed timelines—this tool simplifies post-session analysis, reporting, and workflow auditing without the manual overhead of parsing raw JSON logs. It is particularly useful for developers who need to document their AI-assisted progress or analyze tool-calling patterns across multiple sessions.

Key Features

Advanced filtering to isolate tool usage, subagent interactions, or system messages.
Customizable verbosity levels ranging from minimal summaries to detailed event logs.
Automatic extraction of session metadata including duration, branch, and tool statistics.
0 GitHub stars
Multi-format support for Markdown and machine-readable JSON outputs.
Batch processing capabilities for analyzing entire directories of session transcripts.

Use Cases

Reviewing past AI coding sessions to understand specific implementation steps or tool patterns.
Generating concise reports and documentation of work performed during an automated session.
Auditing AI behavior and subagent performance for quality assurance and workflow optimization.

About

Key Features

Advanced filtering to isolate tool usage, subagent interactions, or system messages.
Customizable verbosity levels ranging from minimal summaries to detailed event logs.
Automatic extraction of session metadata including duration, branch, and tool statistics.
0 GitHub stars
Multi-format support for Markdown and machine-readable JSON outputs.
Batch processing capabilities for analyzing entire directories of session transcripts.

Use Cases

Reviewing past AI coding sessions to understand specific implementation steps or tool patterns.
Generating concise reports and documentation of work performed during an automated session.
Auditing AI behavior and subagent performance for quality assurance and workflow optimization.