Agents & settings
Your AI agents at a glance: detail page, status, AI subscription, accounts, connectors, browser, servers and approvals.
Configure an AI agent: detail page and profile
An AI agent's own page: profile with role text, channels and actions, what you can edit yourself and where the other settings live.
The AI agents page: columns, icons and actions at a glance
What the AI agents list shows: phone number, team member, AI subscription, last activity, status and the icons for Telegram, calls, browser, memories and Manage.
Approvals: when your AI agent asks first
Which steps your AI agent only takes with your approval, how to approve in the chat or on the card, which cards are waiting for you and how to change your agent's limits.
Connect an AI subscription to an AI agent (ChatGPT or Grok)
How to log your AI agent in with your ChatGPT or Grok account: get a code, enter it on the sign-in page, done. With help for error messages.
Email “Your AI agent is working through FlowModels”: what to do now
Why your AI agent is working through FlowModels (formerly FlowAI API), what it costs, how to recognise the fallback and how a new login brings the agent back to its AI subscription.
Multiple accounts, main account and API as fallback
Use several AI accounts for one agent, set the main account and the order, and use API as fallback through FlowModels so your agents never stand still.
Open your AI agent's browser and work along
How to watch your AI agent in its browser: copy the VNC password, fullscreen, the browser's own page in a new tab, tidying up tabs and what helps when access has expired.
Set up MCP connectors for an AI agent
Give your AI agent access to more tools such as GitHub or your own system: add a connector, store a token, switch it on and off.
AI agent status: understand the light and restart the gateway
What green, yellow and red dots mean, where to read the reason and how to restart the gateway of an AI agent that's stuck.
Your AI agents' servers: status, specs and billing
The Servers page explained: specs, number of agents, last login, what each status means, individually managed servers and what the owner sees about billing.