What problem is PeopleBot trying to solve?
Imagine giving an assistant a project today and returning next week. You want it to know which files it used, what happened, and what remains to be done. You also want another assistant to be able to check its work. PeopleBot’s design puts that information in records you can save and inspect. Git keeps the saved versions; PeopleBot helps identify and organize the inputs and results of an agent’s work. An agent is an AI assistant configured to do a particular kind of task. PeopleBot is the framework around that work. It is not itself an AI model.The public version covered here is 0.1.0a4, an early experimental release. The optional file exercises need no model call. The project workflow uses your existing assistants, accounts, and repository access. The package does not yet provide an unattended coding team. Installing it does not start an agent.
Start with the practical workflow
These lessons are operating procedures, not an automatic installer. They distinguish our coordination practice from the public framework’s APIs.
Optional: learn the local mechanics
The hands-on lessons build on one another. Keep using the same tutorial folder. You can close your terminal and return later; the terminal lesson explains how to reopen the right folder.
A few words you will see
For now, focus on a folder, a file, and a saved version. The other concepts will make more sense after you have seen those work.