Documentation

Everything AstraCode does, and how to use it. If you are new, start with the quickstart.

Get started

Install AstraCode and make your first change with it.

Quickstart — Install, sign in, and get from a fresh editor to your first useful change.
Install AstraCode — Download AstraCode for macOS or Linux, sign in, and open your first project.
The chat panel — Agent, Plan, Ask and Debug modes, AstraOne, and the context meter.

What the agent can do, how much it asks before doing it, and how it checks itself.

Agent overview — What the agent is, how a run works, and where to go next.
Approvals and permissions — The three approval modes, what each one asks about, and the commands that always need a human.
Plan mode — Agree the approach before the agent changes any code.
What the agent can do — The tools the agent has: reading, editing, searching, running commands, browsing, and calling your connectors.
Running commands — How the agent runs shell commands, background tasks for dev servers, and the read-only command list in Ask mode.
Browser tools — Let the agent open your running app, read the page and the console, and click through it.
Checkpoints and undo — Every way back: undoing one run, and restoring the whole workspace to an earlier point.
Infrastructure as code — How the agent works on Terraform, OpenTofu, Pulumi, Kubernetes, Helm and Docker without changing anything real you did not approve.
Running several agents at once — Give each agent its own git worktree so they can attempt the same task without colliding.

What the agent knows about your project, and how that is kept from filling up.

How context works — Attaching files with @, what the agent remembers, project rules, and how AstraCode keeps a long run from filling the window.
Rules — Persistent instructions the agent reads before it starts: per project, per file type, or on demand.
Memories — One-sentence facts about your project that the agent keeps between sessions.

One repository or several, and what changes when you open more than one.

Workspaces — Opening a project, opening several, and what a workspace is.
Working across several repositories — How to open more than one repository at once, and how the agent addresses files when you do.

Give the agent tools from other systems through the Model Context Protocol.

Connectors (MCP servers) — Give the agent tools from other systems (issue trackers, databases, internal APIs) through the Model Context Protocol.

Editor extensions, and skills that teach the agent a repeatable procedure.

Plugins and skills — Editor extensions from Open VSX, and skills that teach the agent a repeatable procedure.
Skills — Teach the agent a procedure you want followed the same way every time.
Hooks — Run your own script when the agent is about to do something, and let it approve, block or rewrite the action.
Dictation — Talk to the agent instead of typing: pick a microphone, record, and get text in the composer.

The AstraCode agent in your terminal: a conversation, one task from a script or CI, a branch review, or inside Zed and JetBrains.

AstraCode CLI — Install astracode-agent, sign in, and use the AstraCode agent in a terminal, a CI job, or Zed and JetBrains.

Settings and lookup material.

Settings reference — Every AstraCode setting, what it defaults to, and what changing it does.
Keyboard shortcuts — The shortcuts AstraCode adds. Every other shortcut works as you already know it.
Models — How AstraOne chooses the model for you, and choosing one yourself with your own key.