Skip to content
Wantzel Studio
Agent fleets

Dozens of agents at once, in one overview

Give every open ticket its own agent and they work side by side. One grid shows each of them: what it does, for which ticket, on which AI, and how long it has been at it.

It runs in your own studio. No containers to set up and no cloud orchestration: the fleet is a choice of how many, made on one screen.

one screen, every agent

How a fleet works

  1. Start many at once

    Tick tickets and start an agent for each of them at once. Or launch a number of agents with their roles, or from a plan whose rows you can still change until you press Start.

    Thirty-two agents in one grid

    The agents panel in developer mode with a table of thirty-two agents: twenty-nine running, three done. Each row shows profile, ticket, task and engine; on the right the conversation of the chosen agent.
  2. Follow every one of them

    Each agent is a row: its status, its ticket, its role, what it is doing, its AI and model, its time and its tokens. Click a row and its conversation and steps are beside it. Filter on running, waiting for you, failed or done.

  3. Stop or resume a group

    Stop one agent, or all of them in one click. Agents that got stuck resume where they left off, as a group. Questions that agents ask you come together in one list, most important first.

    Stop all, resume stuck

    The Agents panel: agents running, each with its own Stop button, and above them Stop all (13) and Resume stuck (6).
  4. Several AI providers in one fleet

    One fleet can mix engines. When a Claude subscription nears its usage limit, new work in the fleet goes to another connected AI that still has room, and the row says why.

Read next

it starts with a conversation

See a fleet at work

EUR 200 a month per organisation

Leave your email address and Floris contacts you to plan a conversation and a demo with your own work in mind.

You first get an email to confirm. Unsubscribing is always one click. Privacy