session-cleaner

To convert agent session JSONL logs into markdown transcripts, you can automatically process raw OpenClaw and Clawdbot files to strip out tool calls, system metadata, and operational noise, generating clean, human-readable conversation records.

Removing system metadata and tool calls from OpenClaw session logs is best handled by automated noise stripping, which filters out extraneous operational data and formats the remaining user and assistant conversations into clean markdown.

Yes, you can batch process multiple AI agent session logs by filtering sessions based on specific dates and agent types like Spock or Scotty, allowing you to manage and convert large volumes of records efficiently.

To filter out low-value cron-generated sessions from agent logs, you can apply automated sanitization rules that exclude these automated entries, ensuring your final markdown transcripts only contain meaningful workflow reviews and debugging data.

Yes, the session cleaner works with both OpenClaw and Clawdbot JSONL files, automatically parsing the raw session data from these platforms to extract structured conversations and output standardized markdown transcripts.

When extracting conversations from raw JSONL session logs, limitations include dependency on the specific OpenClaw and Clawdbot formats, meaning logs from unsupported platforms or heavily customized agent structures may not be fully sanitized.