Your agent Your agent, and its AI plan.
The agent is the AI that does the work. It keeps running where it already does, usually your everyday computer, and uses its own computer through ibara. So it can be the agent you already use, or one I help you choose.
Which agents work
Any agent that can use MCP tools, such as Claude Code, Codex, opencode or omp. ibara gives every agent the same tools and the same access rules, so you're never tied to one.
The AI plan
Your agent's AI plan is yours, billed by its provider, not by me. Many owners already pay for one. I'll help you choose one that suits the work, and explain how it's billed before you commit. If you want the AI to be yours too, your agent can run on a local model, on a computer that can handle it.
What the provider sees
The work happens on your computers, but unless your agent runs on a local model, the model runs with your agent's provider, so it sees what it needs to do the job, such as screenshots and task details. Check your provider's terms before working with sensitive information. I'll point out anything worth keeping away from it.
How a job starts
You can ask your agent for a job the way you already talk to it, naming the computer and the work. For regular jobs, we agree how each one starts: when you ask, or on a schedule.
The software's detail
ibara's docs list the tools every agent gets and how to connect one: ibara.app/docs/agents.