What you can do with Agent Fleet
Agent Fleet runs coding agents such as Claude Code and Codex from the Console in your browser. This page shows what you can do with it and which page to read first.
How it works
Section titled “How it works”| What you do | What the agent does |
|---|---|
| Sign in to the Console and start your workspace | — |
| Connect the agent you use, pick a repository and launch a session | Reads code, writes code and runs commands inside your own workspace |
| Answer question, permission and plan cards | Stops where a decision is needed and waits for you |
| Check the result and ask for a commit or a push | Commits and opens a pull request as asked |
Every member has a workspace of their own. Repositories, uncommitted work and sessions live in it, and sessions keep running after you close the browser.
What you can do
Section titled “What you can do”| What you want to do | Read |
|---|---|
| Hand work to an agent, reply to it, stop it | Sessions |
| Clone a repository, check the changes and commit | Code |
| Choose and connect the agents you use | Agents |
| Let agents use AWS, Google Cloud, Jira, MCP servers and more | Integrations |
| Make images and compare them side by side | Images |
| Set up the workspace environment, or recreate it | Workspace |
| Ask questions, translate or summarize without a repository | Chat and memos |
Where to start
Section titled “Where to start”| You | Read first |
|---|---|
| are opening the Console for the first time | Start your first session |
| want the words first | Terms and the big picture |
| want to use only chat and reading, without a terminal | Start here if you don’t use the terminal |
| manage the team’s members and limits | Administering a tenant |
| build and run the deployment | Operating a deployment |
The tables of what can be done and who can do it are in the reference, and what each release changed is in the release history.