Intelligent Terminal
An experimental, open-source fork of Windows Terminal with agents built in. Ask about an error, kick off a task, and keep working without leaving the shell.
Agents where you already work
Agents live in the terminal itself, with your shell context already in hand. Open it and start working.
A real agent in your shell
Run GitHub Copilot CLI or any ACP-compatible agent in a docked pane, with your shell output already in context.
Errors caught as they happen
A failed command is picked up the moment it runs and handed to your agent with the full context, ready to explain it or suggest a fix.
Run against a local model
Bring your own provider, or point the agent at a model running on your own PC so nothing has to leave the machine.
What's available today
Intelligent Terminal is experimental and ships often, so this list keeps growing. Here is what you get right now.
A pair programmer in the shell
A docked, configurable pane running your agent CLI, with always-ready context on your shell output. Ask about what just ran without pasting it anywhere.
Fix errors automatically
A failed command is caught the moment it happens and handed to your agent with the full error context. Turn on automatic suggestions and the fix is sitting there waiting for your approval.
Bring your own agent and model
Pick GitHub Copilot CLI or any ACP-compatible agent, then point it at whichever model you want, including one running on your own machine. A local Ollama endpoint takes a base URL and a model name, and no API key at all.
Start a task without blocking your shell
Type a question mark followed by your prompt. The terminal adds context from the active pane and runs the agent in a background tab.Source, releases, news, and feedback
Read the code, grab the latest build, follow what the team ships next, or tell them what is missing.