# Run Grok CLI in parallel — Tallos

> Tallos runs Grok CLI, xAI's coding agent for the terminal, as parallel sessions on macOS and Windows. You install it and sign in from Settings → Connections, then every session gets its own git worktree and opens in a chat view with model and reasoning-effort pickers. Tallos uses your existing Grok login or xAI API key.

- Canonical: https://runtallos.com/integrations/grok-cli
- Português: https://runtallos.com/pt/integracoes/grok-cli.md
- Updated: 2026-09-29
- Section: Integrations

## Key facts

- Install and sign in from Settings → Connections
- Chat view with model and effort pickers
- Model list read from your signed-in Grok CLI
- One isolated git worktree per session

## What is Grok CLI?

**Grok CLI** is xAI's AI coding agent for the terminal, run with the `grok` command. It works on the project in your current folder — reading code, editing files and running commands — and signs in with your Grok account or uses an xAI API key.

**Tallos** is a desktop [agent development environment](https://runtallos.com/what-is-an-ade) that installs, signs in and launches the real Grok CLI, runs many sessions of it in isolated workspaces, and puts it next to Claude Code, Codex and your other agents.

## How do I set up Grok CLI in Tallos?

Open **Settings → Connections**, install Grok CLI from its card, sign in, and set it as your default agent. An existing `grok` on your PATH is detected automatically.

1. **Install** — Click **Install**. Tallos shows the official command before running it: `npm install -g @xai-official/grok` or `curl -fsSL https://x.ai/cli/install.sh | bash` on macOS, and `npm install -g @xai-official/grok` on Windows (Node.js required).
2. **Sign in** — Click **Sign in** to run `grok login` in a terminal. Or click **Use API key**: Tallos keeps an `XAI_API_KEY` in your computer's secure storage and passes it only to Grok CLI each time it starts it.
3. **Check the models** — Open **Models** on the card. Tallos reads the list from your signed-in Grok CLI, so the list matches what the CLI offers you.
4. **Make it the default and start** — In **Settings → Agents**, set Grok as the **Default agent**. Add a repository, click **+** for a workspace and describe the task; Tallos starts `grok` in a fresh worktree with your prompt.

> **Permissions** — Tallos starts it in manual mode by default, so it asks before running commands. If you turn on **Let agents run commands without asking** (Advanced, on the welcome screen) or choose the autonomous mode in **Settings → Agents → Agent Permissions**, Tallos launches it with `--permission-mode bypassPermissions`. Worktrees isolate files per task; they are not a security sandbox.

Setting up several agents at once? Follow [connect your agents](https://runtallos.com/learn/connect-your-agents).

## How do I run multiple Grok CLI sessions in parallel?

Create one workspace per task. Each workspace is a separate git worktree with its own branch, files and terminals, so parallel Grok sessions never write to the same working directory.

- **Watch them together.** Split panes in any direction; Tallos's Grok hooks report working, waiting and done in the sidebar.
- **Prompts arrive intact.** Tallos passes your task to `grok` as an argument, after a `--` separator, so a prompt that starts with a word like “help” is not read as a Grok command.
- **Review per branch.** Each workspace has its own diff; comment on lines and send the feedback back to that session.
- **Resume later.** Grok sessions appear in Agent Session History; **Resume** runs `grok --resume <id>` in the right folder.

The isolation model is explained in [parallel workspaces](https://runtallos.com/features/parallel-workspaces).

> Connect Grok CLI and run it in parallel. → https://runtallos.com/signup

## Does Grok CLI open in a chat view or a terminal?

In a chat view, by default, for sessions on your computer. Tallos reads Grok's transcript and shows messages and tool calls for reading, while the Grok CLI keeps running underneath.

- **Model pill.** Choose the model before the first message (passed as `-m` at launch) or change it mid-session (sent as `/model`). The list comes from your signed-in Grok CLI.
- **Effort pill.** Set reasoning effort at launch (`--reasoning-effort`) or mid-session (`/effort`). The choices follow the selected model.
- **View terminal** / **Back to chat** switch to the raw Grok interface and back at any time.
- Want the terminal every time? Set **Settings → Agents → Agent View** to **Terminal**.

Read more in [native chat](https://runtallos.com/features/native-chat).

## How do I use Grok CLI alongside other agents?

Add Grok to a [squad](https://runtallos.com/features/squads) — a leader agent plus members — or run it in its own workspaces next to other agents. In a squad, each slot picks its own agent, model and effort, and Tallos applies Grok's model and effort when it launches that member.

- **One of three attempts.** In *3 attempts, pick the best*, Grok can try the same objective as Claude Code and Codex; the leader recommends one result.
- **Builder or reviewer.** In *Build + review*, Grok can build while another agent reviews, or the other way round.
- **Leader.** Any detected agent can lead, including Grok. It plans, dispatches and reports within your limits: 5 rounds and 4 members at once by default, up to 20 rounds and 12 members.

Every member works in its own worktree, and nothing merges without your approval.

## What does Tallos add to Grok CLI?

| Capability | Grok CLI on its own | Grok CLI in Tallos |
| --- | --- | --- |
| Parallel sessions | Open more terminals and set up git worktrees yourself | One workspace per task, each with its own worktree and branch |
| Install and sign-in | Run the install and `grok login` yourself | Install, sign in or save an API key from Settings → Connections |
| API key handling | Export `XAI_API_KEY` in your shell | Kept in secure storage, passed only to Grok CLI |
| Interface | Terminal UI | Chat view with model and effort pickers, terminal one click away |
| Session status | Visible in its own terminal | Working / waiting / done across all sessions in the sidebar |
| Past sessions | `grok --resume` from the right folder | Searchable history across agents, one-click Resume |
| Other agents | Separate tools | Same app; squads with a leader, members and best-of-N |

_Grok CLI runs as xAI ships it. Tallos adds setup, chat, workspaces and coordination._

## Frequently asked questions

### Do I need a separate subscription to use Grok CLI in Tallos?

No. Tallos uses your existing Grok login or xAI API key and does not resell model access.

### How does Tallos know I'm signed in to Grok?

It reads Grok CLI's local session file. When you are signed in, the Grok card in Settings → Connections shows Connected.

### Can I change Grok's model in the middle of a session?

Yes. In the chat view, the model pill sends `/model` to the running session, and the effort pill sends `/effort`.

### Does the Grok chat view work over SSH?

No. The chat view for Grok needs the session's transcript on your own computer, so remote Grok sessions open in the terminal view.

### Can Grok and Claude Code work on the same repository at once?

Yes. Each runs in its own worktree on its own branch, so their changes never collide. Compare the diffs and keep the one you want.

### Does Grok CLI work in Tallos on Windows?

Yes. On Windows 10+, the Connections page installs Grok CLI with npm, so install Node.js first.

---

**Run your agents in parallel with Tallos** — Claude Code, Codex, Gemini and 30+ agents — each in its own workspace, with squads, chat, terminals and review in one app. Uses the subscriptions you already have.

https://runtallos.com/signup (macOS 13+ · Windows 10+)