AI agent orchestration

XTeam Agents

In active development

A multi-agent framework for AI agent orchestration: an agent team is assembled from a YAML description instead of being coded for each process. So far it runs only in our lab.

The problem

Multi-agent automation for a specific process is expensive to build and change if every edit means rewriting code.

Why the usual tools fall short

Pipelines hard-wired into code do not carry over to a new process without rework, and the make-up and links of the agents cannot be changed with one edit to a description.

How it works

A LangGraph execution graph is built from the YAML description, and the map of links between agents is drawn automatically as a diagram. The agent registry lets you look agents up by role and by what they can do.

Demo

Graph diagram: __start__ → analyze → plan → execute → validate, conditional transitions to commit, plan or reflect, then commit → reflect → __end__; at the side, a “node → agent” table.
The cognitive_os pipeline is assembled from a YAML description into a LangGraph graph. The program drew the diagram itself from the built graph; nobody drew it by hand.

Technology

All our products and libraries →

Where it fits

Automation shaped to a process

An agent team for a specific process, without rewriting code for every task.

The agent line-up lives in a description

The agents and their links change with an edit to YAML, not to code, and the process moves to a new environment with the same description.

What backs this up

The agent map is built from a description: YAML → LangGraph graph → diagram

A working graph is assembled from the YAML description, and the program draws its diagram itself. The registry holds 29 agent roles, searchable by role and skill. We show the diagram and the roles; the platform’s code is closed.

Integration

The agent line-up is described in YAML and assembled into a working graph for the process. The platform is not yet used in products or by clients.

Work it out on your own figures

The AVA calculator works out whether automating a process pays off, on your figures, with no form and no price from us. If it does not pay off, it says so.

Estimate the payback

Tell us about your problem

Describe it in your own words. If it cannot be solved or will not pay off, we will tell you before any work starts.

In the reply we will say what we understood, what we need to clarify and whether it is worth doing.

or email us directly: hello@xteam.pro

New write-ups on Telegram