Prepare the environment
Follow the documented Linux setup, install rootless Podman and provide the required network access and storage.
See how each conversation is isolatedChannelGate AI OS is a company-governed AI operating system that coordinates your team’s AI agents, shared work and access.
Self-hosted in your environment
Put the foundation where your company operates.
Run ChannelGate on an approved Linux cloud or server environment. Connect your chat platform and agent accounts, then give each conversation its own persistent runtime and workspace.
Linux + systemdRootless PodmanYour approved cloud or server
Deployment
A foundation your operators can own.
Connect chat and agents
Configure Slack, or evaluate Teams and Google Chat in Beta. Connect the selected Claude Code or Codex account and approve the first users.
Create the first workspaces
Each conversation gets a container with persistent home storage and approved workspace mounts. Instructions, skills and files give it a defined purpose.
Operate and improve
Inspect health, queues, runs and failures. Manual updates are available in the core product; managed updates require Enterprise. Keep a tested recovery process. Makeitfuture can implement the foundation and train your team.
Explore implementationIllustrative workflow
From setup to first useful work
Choose one bounded use case before widening access across the organization.
Prepare
Agree the environment, process owner, provider accounts and approved system access.
Configure
Install ChannelGate, connect chat and agents, then set channel policy and skills.
Validate
Run the chosen use case, review the result and confirm ownership before expanding.
Frequently asked questions
Deployment questions
Plan where ChannelGate runs and who looks after it.
Where can we run ChannelGate?
ChannelGate runs on a Linux server or cloud instance your organization approves. The documented setup uses Linux with systemd, Node.js and rootless Podman, plus access to your chosen chat platform and AI provider. Your operator prepares the host using the installation guide.
Does self-hosting mean no data leaves our environment?
No. Working files and operational state live in your deployment, but the selected AI provider receives prompt and context. Your chat platform and enabled connected services also process the requests sent to them. Review the documented data flows alongside your providers' terms before choosing which information agents can access.
Who maintains the deployment?
Your operator looks after the host, credentials, backups and recovery. Manual software updates are available in every edition; Enterprise adds administrator-initiated managed updates. Follow the installation and update guidance. Makeitfuture can help establish the operating setup through a separate implementation engagement.
Can we start with one workflow?
Yes. Start with a defined use case, a process owner and approved access to the systems involved. Configure the channel's permissions, instructions and skills, then review the results before extending the setup to more teams. The example rollout on this page follows that sequence; the governance controls define what the agent can do.
Put it to work
Start with one useful workflow.
Explore the setup yourself, or build the foundation alongside Makeitfuture.