dataclaw helps you save your conversations with AI tools like Claude Code, Codex, Gemini CLI, OpenCode, and OpenClaw. It takes your chat logs and turns them into data you can use and share safely.
It removes any private details like secrets and personal info before saving the data. You can then upload the cleaned dataset to Hugging Face, a popular place to share and explore datasets.
This tool is aimed at making your AI chat history accessible and useful without needing technical skills.
To run dataclaw on Windows, make sure your computer meets these needs:
- Windows 10 or later (64-bit recommended)
- At least 4 GB of RAM
- Around 500 MB of free disk space
- Internet connection (for uploading datasets)
- You do not need to install additional software beforehand
Click the big button below to visit the official GitHub page where you can find the latest version of dataclaw.
On that page, look for the latest release or download options. Follow the instructions on GitHub to get the program files.
Follow these steps to set up dataclaw:
-
Go to the GitHub page. Use the download button above to get to the dataclaw repository.
-
Find the latest release or available files. Usually under a "Releases" section on the GitHub page.
-
Download the Windows version. This will be an installer or a zipped folder containing the program.
-
If it is an installer file (.exe):
- Double-click the file to run it.
- Follow the setup prompts, accepting the default options is usually fine.
- Finish the install.
-
If it is a zipped folder:
- Right-click the zip file and choose “Extract All.”
- Open the extracted folder.
- Find the main program file (
dataclaw.exeor similar). - Double-click to run.
-
Run dataclaw for the first time. The app may open a setup or welcome screen.
-
Follow any on-screen instructions. You might be asked to select your AI chat logs folder. This is where dataclaw will look for the data to process.
The tool works by taking your saved conversations and cleaning them automatically.
Here is a simple guide:
-
Prepare your chat logs. Before using dataclaw, make sure your AI tools (Claude Code, Codex, etc.) save session history on your computer.
-
Open dataclaw.
-
Point dataclaw to your session logs folder. You can usually do this by clicking a “Browse” button and selecting the correct folder.
-
Start the export process. Press the “Export” button or a similar command inside the app.
-
Wait for dataclaw to process your files. It will remove sensitive info and structure the data.
-
The app uploads the resulting dataset to Hugging Face. This step allows you to share your cleaned chat history with others if you want.
-
Track your uploads by searching for the “dataclaw” tag on Hugging Face.
dataclaw removes private details such as passwords, personal data, and API keys. It works automatically to make sure your exported data is safe to share. You keep full control of what gets uploaded.
Because dataclaw only works from files saved on your computer, it never sends your raw chat history anywhere.
-
Program won’t start:
Make sure you used the correct download file and completed installation steps. Check if your Windows system is updated. -
No chat logs found:
Confirm your AI tool saves local session files. dataclaw needs these to work. -
Export fails or freezes:
Try restarting dataclaw or your PC. Also, check your internet connection for uploading datasets. -
Dataset not uploaded:
Check your internet connection. dataclaw requires access to upload data to Hugging Face. -
Permissions error:
Run dataclaw as an administrator. Right-click the program and select “Run as administrator.”
The data you export using dataclaw joins other datasets tagged with dataclaw on Hugging Face. This collection grows as people share their AI-human coding conversations.
This allows researchers and developers to explore real-world examples of AI tool use in coding projects.
dataclaw runs on Windows with a simple user interface. You do not need experience with coding or command lines to use it.
- Visit the GitHub issues page for dataclaw to report problems or ask questions.
- Look for updates or guides on the GitHub repository.
- Reach out to the repository owner if you need assistance.