Thriving 
PY A platform for running controlled experiments on teams of LLM agents: multi-agent simulations, an event-sourced run log, swap and fork interventions, and post-hoc evaluation metrics.
https://agencyenterprise.github.io/GlossoGen/ ↗// readme
Documentation · Quickstart · Scenarios · Live demo
A platform for running controlled experiments on teams of LLM agents. A scenario puts several agents in a simulated task with their own roles, channels and tools; the platform plays it out in rounds, records every message, tool call and model response, and scores the run with the metrics you pick. The scenarios shipped here study communication under pressure: no agent holds enough information to succeed alone, and in most of them every character sent costs against a per-round budget. Under that pressure agents compress, abbreviate, and invent shorthand.
Everything else lives in the documentation: installing, running and evaluating simulations, the web UI, the fork and swap flows, and how to ship scenarios and metrics in a package of your own.
glossogen needs Python 3.14 and is not on PyPI, so pin a tag from [the releases…