Build a Zero-Server GitHub Bot to Track AI & Open-Source Projects

Artificial Intelligence is moving at an unbelievable speed. Every day, researchers and open-source developers release brand new models, web interfaces, dataset loaders, and deep learning tools on GitHub. Keeping track of all these releases can feel like a full-time job. If you want to stay on top of the latest technology, you need a way to automate this process. But traditionally, running an automation bot meant setting up a virtual server, paying a hosting provider, configuring databases, and maintaining the infrastructure. For a personal project, that is often too much work and too expensive.

What if you could build a bot that runs entirely for free, does not require a server, and uses a default configuration that takes less than ten minutes to set up? In this guide, we will build exactly that: a serverless GitHub bot. Every 30 minutes, this bot will search the GitHub API for new repositories related to AI and Open Source. It will compile the results, format them into a clean layout, and email them directly to you. We achieve this by running a Python script on GitHub Actions and leveraging GitHub's built-in email notification system.

Code URL: https://github.com/ramsharma12345/automate_blogPost

The Architecture: Serverless and Secret-Free

Before we look at the code, let's talk about how this bot sends emails without an email server. Usually, to send an email from a script, you have to sign up for an SMTP provider (like SendGrid or Mailgun) or use your personal Gmail account. This requires generating API keys, storing them securely in your repository, and managing sender quotas. If your keys leak, spammers could abuse your account.

To avoid this headache, our architecture uses a clever trick: GitHub Issues. In any GitHub repository, when a new issue is created, GitHub automatically sends an email notification to the repository owner and anyone watching the project. Instead of emailing you directly, our bot simply creates a new issue in your private repository containing the list of new AI projects. GitHub's own notification system does the work of formatting and delivering the email straight to your inbox. This means you do not need to configure any SMTP secrets, and the setup is completely default and secure.

Step 1: The Python Automation Script (main.py)

The heart of our project is a lightweight Python script. It handles two simple tasks: fetching the data from the GitHub API and posting that data back as an issue. We use the standard Python library along with the requests library to handle HTTP requests.

To make sure we only get the absolute newest projects, the script dynamically calculates the timestamp for 30 minutes ago. It then requests a list of repositories created after that timestamp that match the keywords 'ai' and 'open-source'. The URL query looks like this:

https://api.github.com/search/repositories?q=ai+open-source+created:>YYYY-MM-DDTHH:MM:SSZ

Once the data arrives, the script loops through the list of repositories. It reads the name, description, star count, and primary language, formatting everything into clean Markdown text. Finally, it makes a POST request to the repository's issues endpoint using the GITHUB_TOKEN provided by GitHub Actions to create the new issue.

Step 2: Scheduling with GitHub Actions

To run this script automatically every 30 minutes, we write a YAML configuration file inside the .github/workflows/ directory. GitHub Actions is a built-in automation platform that lets you run workflows triggered by events or schedules. By defining a cron schedule, we instruct GitHub's cloud servers to spin up a virtual machine, download our code, install Python, and execute our script.

The schedule uses standard cron syntax: */30 * * * * (which translates to 'every 30 minutes'). We also include a workflow_dispatch trigger, which gives us a button in the GitHub UI to run the script manually whenever we want to test it. We also declare the required permissions so that our workflow is allowed to write issues to the repository.

When you first push this folder structure to a brand new repository, GitHub Actions won't show any history. It will display a default setup screen until the workflow file is successfully detected on the default branch:

Step 3: Granting Permissions and Testing

By default, GitHub Actions scripts run in a highly secure environment where they can only read the repository, not modify it. If you try to run the workflow immediately, it will fail when trying to create the issue because of a 'Permission Denied' error. To fix this, you need to adjust one setting in your repository:

·        1. Navigate to your repository page on GitHub.

·        2. Click on the Settings tab at the top of the page.

·        3. In the left sidebar, click Actions, and then select General.

·        4. Scroll all the way down to the Workflow permissions section.

·        5. Change the option from 'Read repository content and packages permissions' to 'Read and write permissions'.

·        6. Click Save.

With the correct permissions, you can head back to the Actions tab, select your workflow, and click Run workflow. GitHub will spin up a runner and execute the task.

Once the run finishes successfully, a green checkmark will appear next to the job, showing that our Python script executed successfully:

Step 4: Inspecting the Results

Now, navigate to the Issues tab in your repository. You will see a newly opened issue. The title will display the current date and time of the search, and the body will list the repositories that were created or updated in the last 30 minutes.

Here is an example of what the generated issue looks like, including the repository link, description, stars, and language format:

Because you are the owner of this repository, GitHub will immediately forward this issue as an email directly to your account. This means you do not need to keep checking GitHub—your email inbox will become your personal dashboard for the latest AI projects!

How to Customize Your Bot

The beauty of this project is its flexibility. Because it is a simple Python script, you can easily modify it to fit your needs. Here are a few ideas on how to customize your search parameters:

·        Change the Keywords: If you are interested in Web Development or Cybersecurity instead of AI, simply open main.py and change the query string from 'ai open-source' to 'react tailwind' or 'security python'.

·        Filter by Stars: If you only want to see high-quality repositories that are already gaining traction, add the stars filter to the search query, like 'ai open-source stars:>50'.

·        Change the Interval: If receiving an email every 30 minutes is too frequent, open schedule.yml and change the cron schedule. For example, '0 0 * * *' will run the bot once a day at midnight, compiling all projects created in the last 24 hours (just make sure to adjust the time threshold in main.py to 24 hours as well!).

Conclusion

By building this project, you have created a powerful, zero-cost, and low-maintenance tool. You did not have to configure an email server, purchase cloud hosting, or manage complex infrastructure. By using GitHub Actions and GitHub Issues, you let GitHub handle the computing power, scheduling, and email delivery for you. You can now relax and watch the latest AI innovations deliver themselves straight to your inbox. Happy coding!