Skip to main content
PeopleBot helps keep track of an AI assistant’s work: the exact files it used, what it tried, and what it saved for later. Its goal is to make work easier to inspect and continue across sessions. It uses Git, a tool that saves versions of files. You do not need to understand all of Git to begin. Our tutorials explain how to plan work with an assistant, give a local agent a clear assignment, verify its report, and continue later from saved evidence. Optional local-file exercises explain the mechanics. New here? Start the beginner learning path. Start with the same human-guided coordination pattern used to develop PeopleBot, then explore the underlying tools.
These guides describe the public 0.1.0a4 experimental framework release. It includes deterministic utilities, bounded review/development components, Git messaging, finite work-cycle code, and usage reporting. Installing it does not start an autonomous worker. See the capability table for the release boundary.

​
Choose your starting point

​
What a durable task needs

A useful task needs a clear objective, exact inputs, an execution boundary, and evidence of the result. For example, a review can refer to one committed parser file instead of whichever version happens to be on a branch when someone reads the report. That record makes later inspection possible. It does not make a model’s answer automatically correct: tests, cited evidence, and independent review still matter.

​
Public distribution

The public repository contains distributable source and release artifacts. These guides require no private development or operational access. Read the release guide to distinguish framework versions from separately released services.

​
License and credit

Guthrie E Services, LLC created PeopleBot. PeopleBot is distributed under GPL-3.0-only. The complete license and licensing declaration accompany the distribution.