All Collections
Beta Features
Introducing Github Sync for CodeGPT (Beta)
Introducing Github Sync for CodeGPT (Beta)

We're excited to announce a new beta feature - Github Sync!

Gustavo avatar
Written by Gustavo
Updated over a week ago

We're excited to announce a new beta feature - Github Sync! This allows you to automatically sync your public Github repositories with CodeGPT to generate training data.


  • To use this service, you must have a paid plan.

  • Repo must be public during the sync process

  • Repo size limit: 100MB total, 5MB for extracted text files

  • Only text files will be extracted

Sync Process

  1. Paste your Github Repo in the CodeGPT document section

  2. Click the sync button

  3. CodeGPT will clone your repo and extract text files

  4. Text files will be compiled into a training document

  5. Training document will be available after sync completes

You can then use this document to retrain your CodeGPT agent. The sync process is manual, so you'll need to re-initiate it whenever you want to pull the latest updates from your repo.

Some things to keep in mind:

  • Private repos are not supported

  • Non-text files like images, PDFs, etc will be ignored

  • If sync fails, check for non-text files or remove files to reduce repo size

  • This is a beta feature, so you may experience issues

Let us know if you have any other questions! We're excited for you to try out Github Sync and would love to hear your feedback.

Here's how it works:

File Types. We only process text files. Certain files, such as those without extensions, Adobe files, music files, and executable files, are not included in the sync. This is to keep the repository optimized and running as smoothly as possible.

Repository Size. We divide this into two stages. The first is the condition of the original repository's weight with all its documents. The maximum size for your repository is 100mb. If your repository meets this condition, our bot only extracts plain text files, these files have a limit of up to 5 mb. If the repository does not meet this condition, the system will send an error message.

Repository Visibility. Additionally, at the moment, the repository must be public to sync. Some users change the visibility of their projects during the sync and then restore privacy. In other cases, you can create a new public project without sensitive information.

Start Training. After the synchronization process is successfully completed, you can initiate the training of your agent by opening the newly created file upon application refresh.

Errors. If we find documents that we cannot process, we ask you to remove unsupported documents. You will see the following error: "Oops! Something went wrong. Maybe the repo has a file that cannot be processed."

Some ideas

Request updated documentation: You can ask the repository about the current documentation and if there are clear usage examples.

Ask technical questions: If you have doubts about how to use a library or how to implement a certain functionality, you can ask questions directly in the repository.

Report issues: If you find errors or problems in the code, you can report them to the repository and provide details on how to replicate them.

Suggest improvements: If you have ideas for improving the code, you can propose enhancements and offer concrete solutions.

Request features: You can suggest new features that you consider useful for the project and explain why they could be beneficial.

Seek guidance on repository usage: You can inquire about how to effectively use the repository or how to configure certain features.

Discuss security issues: If you discover potential vulnerabilities in the code, it is essential to report them immediately to the repository to address them.

By communicating directly with repositories on GitHub, you can enhance collaboration and communication in the world of software development, which often leads to better code and a stronger community.


We do not have permissions to read private repositories or write permissions on your GitHub account. The synchronization is manual, so you will need to do one each time you make changes to your code.

This is a beta service, which means that it is offered as a test and may not function as expected. Our team is continuously working on improving the service, so expect changes to occur quickly. You may experience issues while using this service, and by using it you accept the risk associated with beta software.

By using and accessing the CodeGPT Plus Github Repo Sync (Beta), you agree to be bound by these Terms of Service. If you do not agree to abide by these Terms of Service do not use this beta service.

Did this answer your question?