How to set up Ollama Cloud in Sculptor — step-by-step for beginners

You do not need to install Ollama on your laptop to use Ollama Cloud with Sculptor. You need an Ollama account, an API key, and five minutes in Settings. This guide assumes you have never done it before.

What you are setting up (two ideas only)

Term	Plain meaning
Provider	Which AI backend Sculptor uses for replies (Ollama Cloud, OpenRouter, OpenAI, etc.).
Model	Which brain on that provider (for example a specific cloud model id from your account).

Ollama (local) is different: that is an Ollama app running on your computer (http://127.0.0.1:11434). Ollama (cloud) uses Ollama’s hosted API at https://ollama.com by default. This article is cloud only.

Before you start

• A Sculptor account (register at Sculptor if you do not have one).
• A modern browser (Chrome, Firefox, Safari, or Edge).
• An Ollama account (free to create at ollama.com).

You do not need Docker, a terminal, or a GPU on your machine for Ollama Cloud.

Step 1 — Create an Ollama account

1. Open https://ollama.com.
2. Sign up or sign in.
3. If Ollama offers a Cloud or subscription plan for the models you want, follow their site—some models require a paid tier. Sculptor only needs a valid API key that can list and run models on your account.

Step 2 — Create an API key (bearer token)

1. Go to ollama.com/settings/keys.
2. Click to create a new API key.
3. Copy the key immediately and store it somewhere safe (password manager or temporary note).
   You will not be able to see the full key again after you leave the page.
4. Treat this key like a password: do not paste it into public chats, screenshots, or email.

In Sculptor this value is called Bearer token (optional) on the Ollama Cloud tab.

Step 3 — Open Sculptor Settings

1. Sign in to Sculptor.
2. Open Settings from the app menu (or go to /settings in the browser).
3. Find the LLM provider area on the left: you will see tabs such as Ollama (local), Ollama (cloud), OpenRouter, and others.

Step 4 — Select the Ollama (cloud) tab

1. Click Ollama (cloud) in the provider list.
   That also sets Ollama (cloud) as your active provider for new work (the tab you select is the one Sculptor will use after you save).

2. Read the short hint on the page: Sculptor uses an OpenAI-compatible API. The default base URL is:

   https://ollama.com

   Leave that default unless Ollama’s documentation tells you to use a different URL.

Step 5 — Paste your API key

1. In Bearer token (optional), paste the key you copied from Ollama.
2. The field is a password-style box (you can use the eye icon to show/hide what you typed).
3. You do not need to click Save yet to try listing models, but you must save before you rely on the key in a long coaching session.

If you already saved a key before, the placeholder may say Leave blank to keep saved token—paste a new key only when you want to replace the old one.

Step 6 — List models on your account

1. Click List cloud models.
   Sculptor calls Ollama’s model list (GET /v1/models) using your key.
2. Wait a few seconds. You should see a success message with how many models were found.
3. Open the dropdown Models on your account and pick one (for example a smaller/faster model for your first test).
4. If the dropdown is empty, type a model id in Model id (edit or type custom)—copy the exact id from Ollama’s cloud documentation or your account dashboard.

Tip: If listing fails, check that the key is fresh, that you pasted it with no extra spaces, and that your Ollama account can use cloud models at all.

Step 7 — Save settings

1. Scroll to Save settings (top or bottom of the Settings page, depending on layout).
2. Click Save settings.
3. You should see a confirmation that settings were saved and secrets stored encoded in local storage on this device.

The token field clears after save—that is normal. Sculptor keeps an encoded copy; it is not stored as plain text in the form.

Step 8 — Confirm in the workspace

1. Open or create a project in the workspace (/app).
2. Open Chat.
3. In the composer, check the provider / model control if visible—it should reflect Ollama (cloud) and the model you chose (or your Settings defaults).
4. Send a simple message, for example: “Say hello in one sentence and tell me which model you are using.”

If you get a reply, Ollama Cloud is working.

Step 9 — Optional: per-conversation model

Settings define your default provider and model for new conversations. A single thread can use a different model from the chat composer (provider pill / conversation LLM dialog). That is useful when you want a stronger model only for one strategy phase—without changing global Settings.

Using organisation (team) keys instead

If your company uses organisation defaults:

1. An admin may save a shared Ollama Cloud key under Company account → organisation LLM keys.
2. In Settings, you can choose Prefer organisation defaults for credentials (wording may vary; see the LLM credential section).
3. When organisation mode is active, personal provider tabs may be hidden until you switch back to Use my keys on this device.

Ask your admin if cloud models fail and you expected a company key to apply.

Troubleshooting (rookies)

Problem	What to try
List cloud models fails	New key from ollama.com/settings/keys; paste again; click Save settings; retry.
Unauthorized / API key errors in chat	Open Settings → confirm Ollama (cloud) tab is selected → Save settings again.
Dropdown empty but list succeeded	Type the model id manually in Model id; some accounts return ids that are not in the dropdown yet.
Works on one browser, not another	Settings are per browser/device—repeat setup on each device you use.
Expected local Ollama	You chose Ollama (cloud). Local uses Ollama (local) tab and a running Ollama app on your machine—different path.
Model needs paid tier	Check ollama.com/cloud for which models your subscription includes.

What not to worry about yet

• MCP tools — optional; separate from getting basic chat working. See Connect MCP servers to Sculptor when you are ready.
• Agentic runs — use the same provider defaults once Settings are saved; start a Strategy Pack run from the workspace when your model replies in Chat.

What to read next

• From specialist chat to a library-ready Strategy Pack
• Strategy Pack phases guide
• AI Skills in Sculptor — custom instructions on top of your provider

Keywords: Ollama Cloud setup, Sculptor Settings tutorial, Ollama API key, cloud LLM provider, beginner AI coach setup.