Get started

Before you begin

Updated 6 Nov 20263 min read

You can sign up and explore autobotAI in a few minutes. To run real agents against your environment, have these ready.

1. An autobotAI account

Sign up at autobot.live and verify your email. A free trial lets you explore the platform before you choose a plan.

2. A deployment choice

OptionYou need
SaaSNothing else. Your environment is ready when you sign in.
Self-hosted workspaceAn AWS account and an IAM user or role with administrator permissions to launch a CloudFormation stack, in a supported region.

Not sure which to choose? Compare them in Deployment and models.

3. A plan for where your systems are

autobotAI is hybrid: the platform runs in one place, SaaS or your AWS account, and its agents reach systems in every cloud and data center from there.

autobotAIOne control planeSaaS, or a self-hosted workspace in your AWS accountAgentsApprovalsAudit trailRisk register
Other AWS accountsaccounts, Organizations, VPCsAPI · cross-account role
Azuresubscriptions and VNetsAPI · app registrationAgent in the VNet
Google Cloudprojects and VPCsAPI · service accountAgent in the VPC
SaaS toolsidentity, EDR, SIEM, ITSM, chat, codeAPI · scoped credentials
Your data centersservers, directory, internal appsLinux agentVPN or Direct Connect
KubernetesEKS, AKS, GKE or your ownKubernetes agent

Note which of your systems are private before you connect them:

Your system isWhat you need
Reachable over the internet, such as cloud APIs and SaaS toolsCredentials only. If a tool accepts API calls only from known addresses, allow-list the address autobotAI calls from
Inside a private network, for defensive actionsA Linux host (Ubuntu 22.04 or later) or a Kubernetes cluster in that network, with outbound HTTPS. No inbound ports
An internet-facing app you want penetration-testedA verified domain and one temporary allow-list rule
An internal app you want penetration-testedA self-hosted workspace connected to that network by VPN or Direct Connect. Plan its address range before you deploy

Hybrid and private networks explains each path and what your network team sets up.

4. A model provider

Agents and AI nodes need at least one model. Have credentials ready for one of:

5. Access to your tools

Decide which tools your first agents need and get credentials with the right scope:

  • Read-only access for tools agents only investigate, such as threat intelligence, SIEM or vulnerability scanners.
  • Write access only for actions you plan to approve, such as a cloud account, firewall or code repository.
  • A messaging tool such as Slack or Microsoft Teams, so approvers can decide where they already work.

Each integration guide lists the minimum permissions it needs.

6. Your first use case

Pick one risk that costs your team time every week, such as phishing reports, exposed access keys, public storage buckets or access reviews. You'll build it in the Quickstart.

Helpful background

You don't need to write code. It helps to know three ideas:

  • Triggers start a workflow: a schedule, an event or a request.
  • Actions do the work in your tools.
  • Approvals pause the work until a named person decides.

Agents, workflows and harnesses explains the rest.

Need a hand with this?Our engineers can walk through it with you on your own stack.