Snapshots for AI is a devops tool written in Python designed to quickly generate a machine-readable overview of your project so you can more efficiently pass code-related updates over to your favorite LLM.
This asset was developed for a client who wanted to be able to make changes to their code base across multiple files inside their IDE and then inform the LLM of their changes.
This would help the two align better for future prompts. To be frank, we also wanted this one for ourselves, too. We use it ourselves in our open source projects.
How to use Snapshot.py
After dropping the files into your project, open up a terminal in your project’s root directory and run
py snapshot.py and then a markdown file will be generated inside a
With LLMs like ChatGPT, there is currently a limit to the number of characters you can pass to the model as a prompt so we allow the
snapshot.py file to be edited directly to defined patterns to ignore or focus on to help reduce the size of the markdown file. We also allow the user to defined a custom message that is prepended to the markdown file, to help with the prompt to the LLM
Here’s the asset’s features broken down:
Features of Snapshots.py
- Auto detect sensitive keys in JSON and obfuscate them to better protect your data from AI.
- Customizable pattern-based ignore system that helps prevent the inclusion of unwanted information in your markdown reports.
- Customizable prependable message on every snapshot that attempts to explain the nature of the markdown document to the AI.
Example of Snapshots.py for AI Output
Here’s an example of the markdown output you can expect from Snapshots for AI. Let’s say you’re working on a simple entry-level coding application, and you run the Snapshots for AI tool.
Here’s what a project markdown snapshot looks like for a small project:
Snapshots.py for AI is available for free on Github. All we ask is that you star the repository if you use it.
If you like our content, please consider following us! If you like our free open source assets, please give them a github star. We’re also happy to have lurkers on our Discord community where we manage our syndication network and curate together.