Quickstart¶
Convert your first ChatGPT export in a few minutes.
Prerequisites¶
- uv package manager
Step 1: Choose how to run the CLI¶
Use one of these tool-first options:
Option A: Run without installing¶
Option B: Install as a global CLI (recommended for frequent use)¶
If you are contributing code in this repository, use the development setup in Installation.
Step 2: Export your ChatGPT data¶
- Go to ChatGPT Settings → Data controls → Export data.
- Click Export and wait for the confirmation email from OpenAI.
- Download the ZIP file from the email link.
Step 3: Run the converter¶
You can pass either the ZIP file directly or an extracted directory.
If you installed the tool globally, run:
Step 4: Explore the output¶
The archive/ directory now contains your conversations:
archive/
├── index.md # Start here
├── conversations/
│ ├── index.md # Table of all conversations
│ └── 2024-08-24_my-first-chat_17cd7535/
│ ├── index.md # Rendered conversation
│ └── media/
│ └── 001-a1b2c3d4.jpeg
├── dalle/
│ └── index.md
└── metadata/
├── user_profile.md
├── settings.md
├── feedback.md
└── export_stats.md
Open archive/index.md in any Markdown viewer, or serve it with a static site generator.
Common options¶
Control what gets included:
# Keep PII in user metadata (redacted by default)
uvx chatgpt-to-markdown ./export ./archive --no-redact-pii
# Include thinking/reasoning blocks from o-series models (excluded by default)
uvx chatgpt-to-markdown ./export ./archive --include-thinking
# Skip SHA-256 deduplication (enabled by default)
uvx chatgpt-to-markdown ./export ./archive --no-deduplicate
If you installed the CLI with uv tool install, remove the uvx prefix in all
examples.
Next steps¶
- Installation — more installation methods and system requirements
- Configuration how-to — tune every aspect of the conversion
- CLI reference — complete list of flags and exit codes
- Output structure reference — understand the archive layout