Bloom Docs
Browser and terminals
Keep a running app, terminal output and the conversation in one workspace. The agent can prepare those panes for you.
Open your app beside the agent
A coding task often needs a running app, a terminal and the conversation at once. Use a browser tab for a page you check occasionally. Use a split when you want the page visible while you talk to the agent.
Useful quick prompts
- Open in tab
- Open the site in a new browser tab. Read the URL from
.env. - Open in split
- Open the site in a browser split beside this chat. Read the URL from
.env. - Start the dev server
- Open a terminal tab named Dev server and run the project run command in it.
Agents inside Bloom can manage these panes through MCP. They can open, rename and split tabs, type into a terminal, navigate the browser, read visible page text and send a screenshot back into the chat. Bloom asks before an action exposes page contents or moves a browser you are using.
Terminals, browsers and panes
Every workspace has a terminal in its own worktree, so npm run dev
in one workspace cannot collide with the same command in another. Turn on
"Keep terminals running after quitting" in Settings and terminal sessions are
handed to tmux instead, so they keep running after you quit Bloom and are still
there when you open it again. That needs tmux installed, and the setting says
so when it is not.
- New terminal tab: ⌘⇧T.
- New browser tab: ⌘⇧B, with the Web Inspector docked inside it.
- Split a pane: ⌘\ to the right, ⌘⇧\ downwards. Close one with ⌘⌃W.
- Rearrange by dragging a tab onto another tab or onto a pane, and a pane by the grips on the divider beside it.
Each workspace is allocated ten ports, and the first is in
BLOOM_PORT, so a run script can bind something that will not
clash with the workspace next to it.