skills.rest
megumi-ben
megumi-benCOMMUNITY
Explore all skills by this author →
GitHub

chatlog

Extract conversation messages from Claude Code or Codex session files into JSON.

Updated Mar 27, 2026
Search Tags:#claude code#codex#json#jsonl#conversation export#chatlog#session parsing
One-click install
npx skills add https://github.com/megumi-ben/work17 --skill chatlog-megumi-ben
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: chatlog
Source: https://github.com/megumi-ben/work17/tree/main/.codex/skills/chatlog
Command: npx skills add https://github.com/megumi-ben/work17 --skill chatlog-megumi-ben

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI coding sessions in Claude Code and Codex are stored as raw JSONL files with mixed record types, making it hard to get a clean copy of the actual conversation. This Skill locates the current session file and converts it into a structured JSON file of user and assistant messages. ## Core Features & Use Cases - Automatic Tool Detection: Checks the CLAUDECODE environment variable to determine whether the session belongs to Claude Code or Codex. - Session File Discovery: Finds the most recent session JSONL file from ~/.claude/projects or ~/.codex/sessions. - Message Extraction: Parses JSONL records, merges consecutive same-role messages, filters non-message records, and writes a clean {"messages": [...]} JSON output. - Use Case: After a long debugging session, ask the assistant to save the chatlog so you can archive the Q&A, share it with teammates, or feed it into documentation. ## Quick Start Ask the assistant to save the current conversation to a chatlog JSON file.

Frequently Asked Questions about chatlog

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I save my Claude Code conversation to a file?▼

Claude Code stores sessions as JSONL files under ~/.claude/projects/<project-slug>/. The extraction script filters user and assistant records, pulls text content, and writes a JSON file with a messages array.

How do I export a Codex session conversation?▼

Codex sessions live in ~/.codex/sessions/YYYY/MM/DD/ as rollout JSONL files. Filter records where type is response_item and payload.type is message, then extract text items from payload.content into JSON.

Where are Claude Code session files stored?▼

They are stored at ~/.claude/projects/<project-slug>/<session-uuid>.jsonl, where the project slug is the working directory path with slashes replaced by dashes. The newest file is the current session.

Why does the exported chatlog merge consecutive messages?▼

Claude Code streams assistant replies as multiple chunks, so consecutive same-role records are merged into one message. This produces a clean alternating user/assistant transcript instead of fragmented entries.

Does the extraction include tool calls and file snapshots?▼

No. Records whose type is not user or assistant, such as file-history-snapshot entries, are skipped. Only plain text content from user and assistant messages is kept in the output.

Related Skills

Explore More Software Engineering Skills→
ElliotDrelElliotDrel
1

estack-vscode-file-recovery

Recover deleted files from editor Local History or Claude session transcripts.

Community
Advanced
HuXiangyu123HuXiangyu123
3

lark-workflow-meeting-summary

Summarize Lark meetings within a date range into structured reports.

Community
Advanced
cattboycattboy

learn

Consolidate, search, prune, and export project learnings across sessions.

Community
Intermediate
OKHP3OKHP3
1

okhp3-thread-extract-claude

Extract pasted Claude conversations into standalone, provenance-aware Markdown artifacts.

Community
Advanced