Skip to content
ExoHuman ExoHuman | Docs

Introduction

ExoHuman is the operating system for your company’s AI. It gives every team a fleet of autonomous agents that run in isolated containers, use real tools, remember what they did, collaborate with each other and ask a person whenever a decision needs a human.

Instead of one chatbot per person, you run a fleet: a research agent that monitors competitors every morning, a finance agent that reconciles invoices, a support agent that answers on Slack and escalates refunds for approval, and a principal assistant that coordinates them on your behalf.

What you get

Section titled “What you get”
  • A dedicated instance. Every customer gets their own ExoHuman instance with its own database, scheduler, MCP server and agent containers. Instances are never shared between customers.
  • Agents as containers. Each agent is a long-lived container with its own workspace files, instructions (AGENTS.md), skills, credentials, schedules and resource limits.
  • A choice of agent runtimes. Agents run on Pi (the default), Claude Code, Codex, Gemini CLI or OpenCode. See Runtimes.
  • Humans in the loop. Agents send approvals, questions and alerts to a shared Inbox and wait for an answer when they need one.
  • Collaboration. Agents delegate work to colleagues they are allowed to talk to, react to each other’s events and can federate with external agents over A2A. See Agent network.
  • Operations built in. Schedules, loops, projects, canvases, an audit log, guardrails, model budgets and an operations dashboard.
  • Open interfaces. A REST API, an MCP server with the whole platform as tools, and a CLI.

How ExoHuman runs

Section titled “How ExoHuman runs”

Every customer gets a dedicated, isolated instance on ExoHuman Cloud — never a shared tenant. Agents run as containers on elastic worker pools in the region you choose, with CPU, high-memory, GPU and custom profiles, scaled per agent and billed per minute of running time.

Machines you own can join the same fleet: any Linux server, VM or GPU machine connects as a worker node in two steps, and agents you place on it run on your hardware, inside your perimeter, under the same governance.

Read Choosing a deployment to see where your agents can run.

Where to go next

Section titled “Where to go next”

© 2026 ExoHuman. The operating system for your company's AI.