Smart Terminal
One terminal space shared by you and your agents. Type in plain English if you like.
Tortilloid has its own terminals, built into the project window. They are one shared terminal space: you can see every terminal your agents are using, and your agents can see yours.
Open a terminal
Press Cmd+T to open a terminal in your project.
Type in plain English
- Type what you want in your own words, for example “show me what's using port 3000”.
- Press Cmd+Enter. Kongo turns it into a proper terminal command.
- Check the command, then press Enter to run it.
Agents work here too
- Agents run their commands in Tortilloid's terminals, so you can watch the work as it happens.
- Agents manage their own terminals. They open what they need and close the ones that are no longer useful.
- If an agent needs you to do something in a terminal, such as sign in or approve a step, it points you to the right one.
- Kongo's commands follow your permissions.
Secure by design
- Agents see what you see. Kongo and connected agents can read and use Tortilloid's terminals, so every command they run is out in the open, never hidden.
- Kongo asks first. Kongo runs commands only with your OK, under the rules you set in Tools and permissions. Sensitive folders such as
.sshalways need your approval. - Started in your project. Agent terminals start in the project folder. That sets where commands begin, it isn't a sandbox, so a command can still reach other folders.
- Connected agents bring their own rules. Claude Code, Codex and other connected agents follow their own approval settings, not Kongo's. See Known limits.
Ask about anything you see
Select output in any terminal and send it to Kongo to talk it through. An error you don't understand, a log that looks wrong, a command you're unsure about. Agents see the same terminals you do, so they can explain what happened and what to do next.
Don't care about terminals?
Then you don't need to look. Agents take care of everything, and the terminals are there whenever you are curious.