For AI agents: The documentation index is at https://docs.digitalocean.com/llms.txt. Markdown versions of pages use the same URL with index.html.md in place of the HTML page (for example, append index.html.md to the directory path instead of opening the HTML document).

What is Grok Build?

Grok Build is a command-line coding agent (with an interactive terminal UI) from xAI. Describe what you want in natural language and Grok plans the work, edits files, runs commands, and shows clean diffs for review. This image wires Grok Build to DigitalOcean Gradient’s OpenAI-compatible inference endpoint, so a single Gradient model access key unlocks GPT-5.5, Claude, Llama, Kimi, GLM, DeepSeek, Qwen, MiniMax and more — plus DigitalOcean’s Intelligent Inference Router.

• Terminal-first – Works natively in your shell; no web interface required
• DigitalOcean Gradient AI – Pre-configured inference with one model access key
• Intelligent Inference Router – Optionally route each prompt to the best-fit model
• Plan mode – Generate a plan, comment on or rewrite steps, then approve before changes
• Parallel subagents – Delegate large tasks to subagents that run in parallel
• Headless mode – Run agents non-interactively in scripts with grok -p
• Works with your conventions – Picks up AGENTS.md, plugins, hooks, skills, and MCP servers

Key Features

• Natural-language, agentic coding assistance in the terminal
• Pre-configured with DigitalOcean Gradient AI (GPT, Claude, Llama, DeepSeek, and more)
• Optional DigitalOcean Intelligent Inference Router (router:<name>)
• Plan / review / approve workflow with clean diffs
• Headless scripting (-p) with plain, json, and streaming-json output
• xAI account sign-in supported as an alternative provider

System Requirements

Grok Build is lightweight; most compute happens at the inference service.

Included System Components

• Ubuntu 24.04 LTS – Base operating system
• Grok Build – xAI terminal coding agent (version 0.2.51)
• DigitalOcean Gradient AI – Pre-configured inference provider
• Git – Version control
• curl, jq, unzip – Utilities
• UFW Firewall – SSH only (rate-limited)

Software Included

This Marketplace listing does not include a detailed software list.

Creating an App using the Control Panel

Click the Deploy to DigitalOcean button to create a Droplet based on this 1-Click App. If you aren’t logged in, this link will prompt you to log in with your DigitalOcean account.

Creating an App using the API

In addition to creating a Droplet from the Grok Build 1-Click App using the control panel, you can also use the DigitalOcean API. As an example, to create a 4GB Grok Build Droplet in the SFO2 region, you can use the following curl command. You need to either save your API access token to an environment variable or substitute it in the command below.

curl -X POST -H 'Content-Type: application/json' \
         -H 'Authorization: Bearer '$TOKEN'' -d \
        '{"name":"choose_a_name","region":"sfo2","size":"s-2vcpu-4gb","image":"grokbuild"}' \
        "https://api.digitalocean.com/v2/droplets"

Getting Started After Deploying Grok Build

Getting Started

1. Deploy the Droplet

1. Select this 1-Click App from the DigitalOcean Marketplace
2. Choose a Droplet size (1 GB RAM minimum)
3. Add your SSH key for secure access
4. Create the Droplet

2. SSH In

ssh root@your-droplet-ip

3. Complete the Setup Wizard

On first login, the setup wizard prompts for your DigitalOcean Gradient model access key. To create one:

1. Go to https://cloud.digitalocean.com/gen-ai
2. Navigate to API Keys > Model Access Keys
3. Click Create Model Access Key and copy it
4. Paste it when prompted, then pick your default model from the list (or choose the Intelligent Inference Router)

The wizard verifies the key against https://inference.do-ai.run/v1 and configures Grok Build automatically. You can change the model later with /model in the TUI or -m <alias>.

Prefer to use xAI directly? Press Enter at the Gradient prompt to choose xAI account sign-in or to enter an xAI API key.

No browser needed. With a Gradient key (or any API key), Grok authenticates with the API directly — there is no OAuth browser step. Because the droplet has no desktop browser, the xAI account path uses device-code sign-in (/opt/grok-login.sh), which shows a short URL and code to open on your laptop or phone. Avoid running a bare grok login, which would try to launch a local browser.

4. Run Grok Build

cd /path/to/your/project
grok

For automation:

grok -p "Explain this codebase"
grok -p "Review this diff" --output-format json --always-approve

The default model is GPT-5.5 (gpt-5-5 via DigitalOcean Gradient). Switch with /model in the TUI or -m <alias> headlessly.

Pre-Configured Models

All models below are available via DigitalOcean Gradient with just a model access key. The setup wizard lets you pick the default from this list (or a router). Switch any time with the alias via -m or /model; configuration lives in /root/.grok/config.toml.

DigitalOcean’s serverless catalog grows over time. List the live set on the droplet with:

curl -s -H "Authorization: Bearer $MODEL_ACCESS_KEY" \
  https://inference.do-ai.run/v1/models | jq -r '.data[].id'

To use a model that isn’t pre-configured, add a [model.<alias>] block to /root/.grok/config.toml (copy an existing Gradient entry and change model to the catalog ID).

Intelligent Inference Router

DigitalOcean’s Inference Router classifies each prompt and sends it to the best-fit model based on rules you define (optimizing for cost or latency).

1. Create a router under Gradient > Inference > Routers in the control panel (or via the API) and attach it to the same model access key you use for the droplet.
2. Set GRADIENT_ROUTER=<router-name> in /opt/grok-build.env (or as a droplet env var) and run /opt/apply-gradient-from-env.sh, or enter the router name in the setup wizard.

This points the router alias at router:<router-name> on https://inference.do-ai.run/v1 and makes it the default. The router authenticates with the same MODEL_ACCESS_KEY as the direct models — no extra key needed. Using the router is a drop-in replacement for a specific model.

Managing Grok Build

Helper Scripts

Configuration

• Grok config: /root/.grok/config.toml
• API key env template: /opt/grok-build.env
• Active key: /etc/profile.d/grok-build-key.sh (MODEL_ACCESS_KEY or XAI_API_KEY)
• Getting started guide: cat /root/grok_build_info.txt

Start, Stop, Restart, and Update

Grok Build is an interactive CLI, not a long-running background service, so there is no daemon to start, stop, or restart.

• Start a session: grok (or grok -p "..." for headless)
• Stop a session: exit the TUI (Ctrl-C / /exit) or end the headless command
• Restart: simply run grok again
• Update: /opt/update-grok-build.sh (re-runs the official installer for the latest stable release)

Troubleshooting

grok not found in PATH

Ensure you’re in a login shell (SSH session) where PATH is set, or run directly:

/root/.grok/bin/grok

AI features not working

Re-run the setup wizard to reconfigure your Gradient model access key:

/opt/setup-grok-build.sh

Or set the key manually and reload your shell:

echo 'export MODEL_ACCESS_KEY="<your key>"' > /etc/profile.d/grok-build-key.sh
chmod 600 /etc/profile.d/grok-build-key.sh

Create keys at https://cloud.digitalocean.com/gen-ai (API Keys > Model Access Keys).

Additional Resources

• Grok Build docs: https://docs.x.ai/build/overview
• Custom models reference: https://docs.x.ai/build/overview#custom-models
• DigitalOcean Inference Router: https://docs.digitalocean.com/products/inference/how-to/use-inference-router/
• Announcement: https://x.ai/news/grok-build-cli

Support

For Grok Build-specific issues:

• Documentation: https://docs.x.ai/build/overview
• Send feedback from inside the CLI with /feedback

For DigitalOcean Droplet issues:

• DigitalOcean Support: https://www.digitalocean.com/support
• Community Tutorials: https://www.digitalocean.com/community

Note: This 1-Click installs Grok Build via the official xAI installer and pre-configures DigitalOcean Gradient AI as the inference provider. SSH is the only exposed port; there is no web interface. You can alternatively authenticate with an xAI account (SuperGrok / X Premium+) or an xAI API key.