FAQ
Product
Section titled “Product”What is ExoHuman?
Section titled “What is ExoHuman?”ExoHuman is the operating system for your company’s AI: a platform where you create, run and govern a fleet of AI agents that work alongside your people — with chat and voice, schedules, approvals, collaboration between agents, credentials and a full audit trail. Start with the Introduction.
Can I download ExoHuman and host it myself?
Section titled “Can I download ExoHuman and host it myself?”ExoHuman runs on ExoHuman Cloud, which we host and operate. You can also connect machines you own as worker nodes, so agents with special needs run on your own hardware inside your perimeter. See Choosing a deployment.
Which AI models and runtimes can agents use?
Section titled “Which AI models and runtimes can agents use?”Agents run on a runtime — Claude Code, Codex, Gemini CLI, OpenCode or Pi — with models from the providers your administrators connect (Anthropic, OpenAI, Google and OpenAI-compatible endpoints). Bring your own Anthropic, OpenAI or Google keys and pay the provider directly with no markup, or use the managed model gateway billed at provider list price plus 5%.
Is the ExoHuman Brain available?
Section titled “Is the ExoHuman Brain available?”Not yet. The ExoHuman Brain is planned; that page explains what exists today.
ExoHuman Cloud
Section titled “ExoHuman Cloud”Is my instance shared with other customers?
Section titled “Is my instance shared with other customers?”No. Every customer gets a dedicated, isolated instance — its own database, secrets, storage and agent containers. See Instances & architecture.
How do I get started?
Section titled “How do I get started?”Sign up at https://exohuman.cc/signup, configure your instance (subdomain, region, plan, default compute profiles), pay with Stripe Checkout and open your URL once provisioning finishes. See Quick start and Sign up & create your instance.
How much does it cost?
Section titled “How much does it cost?”A monthly plan fee plus metered usage. Launch is €49/month and Scale is €390/month; Enterprise is priced by contract. Compute is billed per compute-profile hour, metered per minute, and storage above the included amount per GB-month. Prices exclude VAT. See Plans & pricing and Usage-based billing.
Do I pay for agents that are idle?
Section titled “Do I pay for agents that are idle?”Agents are billed only while a container is running a task. Idle agents are suspended and their memory stays in git. See Autoscaling & worker pools.
Can agents use GPUs?
Section titled “Can agents use GPUs?”Yes. Choose a GPU compute profile per agent, or define a custom profile. GPU profiles are available on the plans listed in the profile table.
Where is my data stored?
Section titled “Where is my data stored?”In the region you choose at sign-up — EU · Frankfurt, EU · Helsinki, US · N. Virginia (default: EU · Frankfurt). See Regions & data residency.
How do I change plan, update my card or download invoices?
Section titled “How do I change plan, update my card or download invoices?”From https://exohuman.cc/account, which opens the Stripe customer portal for payment methods and invoices. See Usage-based billing & invoices and Changing plan.
How do I keep spending under control?
Section titled “How do I keep spending under control?”Set per-team model budgets, per-agent timeouts and parallelism, and keep long-running agents on smaller profiles. See Usage limits & budgets.
What happens if I cancel?
Section titled “What happens if I cancel?”Your instance keeps running until the end of the paid period, then stops; data is retained for a grace period so you can reactivate or export. See Cancelling & data export.
Worker nodes
Section titled “Worker nodes”Can we run agents on our own machines?
Section titled “Can we run agents on our own machines?”Yes. Connect any Linux server, VM or GPU machine as a worker node in two steps from the admin panel. Agents you place on your nodes run as containers on your hardware, under the same approvals, resource limits and audit logging as cloud agents.
Do the machines need to be reachable from the internet?
Section titled “Do the machines need to be reachable from the internet?”No. Nodes connect outbound to the instance; no inbound ports are opened on your machines.
Can it run without internet access?
Section titled “Can it run without internet access?”No. Worker nodes need outbound connectivity to the control plane. Fully disconnected operation is not supported.
Agents
Section titled “Agents”What exactly is an agent?
Section titled “What exactly is an agent?”An agent is a long-lived AI worker with its own container, files, memory in git, credentials, schedules and permissions. It does real work through a runtime and its tools. See Key concepts and Creating agents.
Will agents act without my approval?
Section titled “Will agents act without my approval?”Only within the bounds you set. New agents have autonomy off so nothing runs on a schedule until you enable it, and agents can be required to ask a person before sensitive actions. See Guardrails and Inbox & approvals.
Can agents work together?
Section titled “Can agents work together?”Yes — through delegation and rooms, along permissions you grant, and through event subscriptions.
Can I talk to agents from Slack, Telegram, WhatsApp or the phone?
Section titled “Can I talk to agents from Slack, Telegram, WhatsApp or the phone?”Yes. See Slack, Telegram, WhatsApp and Voice & telephony.
Can I use ExoHuman from Claude Code or Cursor?
Section titled “Can I use ExoHuman from Claude Code or Cursor?”Yes, through the MCP server.
Security
Section titled “Security”Who can see my agents and data?
Section titled “Who can see my agents and data?”Only people your administrators admit, according to their role and what is shared with them. Our operations staff access instances only to operate and support them. See Security & compliance.
Do you train models on my data?
Section titled “Do you train models on my data?”No. ExoHuman does not train models. Model calls go to the providers you connect, under their terms.
Do you support SSO?
Section titled “Do you support SSO?”Yes — OpenID Connect single sign-on, SCIM provisioning and two-factor authentication. See SSO & 2FA.