my-personal-claude-code-data

# Coding Agent Conversation Logs > **This is a performance art project.** Anthropic built their models on the world's freely shared information, then introduced increasingly [dystopian data policies](https://www.anthropic.com/news/detecting-and-preventing-distillation-attacks) to stop anyone else from doing the same — pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share. Exported with [DataClaw](https://github.com/banodoco/dataclaw). **Tag: `dataclaw`** — [Browse all DataClaw datasets](https://huggingface.co/datasets?other=dataclaw) ## Stats | Metric | Value | |--------|-------| | Sessions | 549 | | Projects | 14 | | Input tokens | 15.1B | | Output tokens | 4.6M | | Last updated | 2026-02-24 | ### Models | Model | Sessions | |-------|----------| | claude-opus-4-6 | 270 | | claude-opus-4-5-20251101 | 256 | | claude-sonnet-4-6 | 16 | | claude-haiku-4-5-20251001 | 6 | | claude-sonnet-4-5-20250929 | 1 | ## Schema Each line in `conversations.jsonl` is one conversation session: ```json { "session_id": "uuid", "project": "my-project", "model": "gpt-5.3-codex", "git_branch": "main", "start_time": "2025-01-15T10:00:00+00:00", "end_time": "2025-01-15T10:30:00+00:00", "messages": [ {"role": "user", "content": "Fix the login bug", "timestamp": "..."}, { "role": "assistant", "content": "I'll investigate the login flow.", "thinking": "The user wants me to...", "tool_uses": [{"tool": "Read", "input": "src/auth.py"}], "timestamp": "..." } ], "stats": { "user_messages": 5, "assistant_messages": 8, "tool_uses": 20, "input_tokens": 50000, "output_tokens": 3000 } } ``` ### Privacy - Paths anonymized to project-relative; usernames hashed - No tool outputs — only tool call inputs (summaries) ## Load ```python from datasets import load_dataset ds = load_dataset("peteromallet/dataclaw-peteromallet", split="train") ``` ## Export your own ```bash pip install dataclaw dataclaw ```