I Gave My AI Agents an Org Chart

Description and use

Botropolis models a company of 20 AI agents with a CEO orchestrator, offering transparent coordination, modularity, and vendor‑agnostic tooling without API keys.

Problem

Agent frameworks often require manual wiring and opaque orchestration, making complex systems hard to explain and maintain.

Outcome

Botropolis demonstrates that a company‑structured approach to AI agents yields clear ownership, modularity, and ease of extension, while remaining vendor‑agnostic and free of API keys.

Workflow

• 1. Define agents with YAML specs (name, title, specialty, tools, example tasks).
• 2. Deploy the CEO orchestrator to route requests to relevant departments.
• 3. Agents collaborate via teaming rounds, drafting, critiquing, and revising outputs.
• 4. Use built‑in tools (web search, shell, file, Gmail, calendar, shared notebook) without vendor lock‑in.
• 5. Interact through a web UI or programmatically via the CEO API.
• 6. Optionally train or fine‑tune agents using the provided pipeline.

Evidence

Most agent frameworks hand you a pile of scripts and wish you luck. You wire up three agents, write the orchestration glue yourself, and six weeks later you have a system nobody can explain.

Source

It is called Botropolis. Twenty specialist agents across ten departments, one CEO orchestrator, and a training pipeline so each agent can eventually get its own fine‑tuned model.

Source

Real tools, no vendor lock‑in. Agents with toolkits run a think‑act‑observe loop over a plain‑text protocol. No provider function calling involved, so it works with OpenAI, Anthropic, Gemini, Ollama, or a model you trained yourself.

Source

A web UI with no build step. Plain HTML, CSS, and JS served by the same FastAPI process. Chat with the CEO, a war room for talking to one agent directly with live tool‑call streaming, teaming controls, per‑agent analytics, and a roster browser. It works on a phone.

Source

It runs with zero API keys against an offline stub that labels its output as stub output. Clone it, run python examples/demo.py, and watch the whole company work before spending anything.

Source