How to connect Ollama in The Novelist

Overview

Ollama runs open-source language models on your own computer. The Novelist talks to Ollama over localhost (your machine only) — your workspace files stay in the folder you chose during setup.

You will:

1. Install Ollama for your operating system.
2. Download at least one model (a few GB; one-time per model).
3. Enable the Local AI (Ollama) plugin in The Novelist.
4. Point Settings → AI Provider at http://localhost:11434 with your model name.
5. Click Test connection, then Save.

Estimated time: 15–30 minutes, mostly waiting for the model download.

Requirements

• The Novelist installed — install guide · download
• Disk space: at least 5–25 GB free depending on the model (see table below).
• RAM: 8 GB minimum for small models; 16 GB+ recommended for better quality; 32 GB+ for large models.
• Internet only for the initial Ollama and model download — after that, AI runs offline.

Install Ollama on macOS

Works on Intel and Apple Silicon Macs. Apple Silicon (M1/M2/M3/M4) is faster for local AI.

Step 1 — Download

1. Open ollama.com/download in Safari or Chrome.
2. Click Download for macOS. You get a file named like Ollama-darwin.dmg.

Step 2 — Install

1. Open the downloaded .dmg file.
2. Drag the Ollama icon into your Applications folder.
3. Eject the disk image, then open Ollama from Applications.

Step 3 — First launch & permissions

1. If macOS says the app is from an unidentified developer, open System Settings → Privacy & Security and click Open Anyway for Ollama.
2. Ollama adds a llama icon to the menu bar. When it is running, local AI is available on port 11434.
3. Optional: open Terminal (Applications → Utilities → Terminal) — you will use it to download models in a later step.

Install Ollama on Windows

Requires 64-bit Windows 10 or Windows 11.

Step 1 — Download

1. Visit ollama.com/download.
2. Click Download for Windows (OllamaSetup.exe).

Step 2 — Run the installer

1. Double-click OllamaSetup.exe.
2. If SmartScreen appears, choose More info → Run anyway (Ollama is a widely used open-source tool).
3. Follow the setup wizard. Ollama installs a background service and adds itself to the system tray.

Step 3 — Confirm it is running

1. Look for the Ollama icon in the system tray (bottom-right, near the clock). You may need to click the ^ arrow to show hidden icons.
2. Open PowerShell or Windows Terminal: press Win + X, choose Terminal, or search for "PowerShell" in Start.
3. Run ollama --version. If you see a version number, the CLI is on your PATH.

Install Ollama on Linux

Official one-line install works on most x86_64 and ARM64 distributions (Ubuntu, Fedora, Debian, Arch, etc.).

Option A — Official install script (recommended)

1. Open Terminal.
2. Run:

curl -fsSL https://ollama.com/install.sh | sh

3. The script installs the ollama binary and a systemd user service that starts automatically.
4. Verify: ollama --version

Option B — Manual install

Download the tarball for your architecture from ollama.com/download/linux, extract it, and place the binary on your PATH. See Ollama's documentation for distribution-specific notes.

Start or check the service

# Check status
systemctl --user status ollama

# Start if stopped
systemctl --user start ollama

# Enable on login
systemctl --user enable ollama

Ollama listens on http://127.0.0.1:11434 by default. Test in the browser or with:

curl http://localhost:11434/api/tags

You should get JSON (possibly an empty model list before your first pull).

Choose a free model

Models differ in quality, speed, and RAM use. Names are case-sensitive — type them exactly in The Novelist, including the colon and tag (e.g. qwen3.5:4b).

Browse more models at ollama.com/library.

Download (pull) your model

Open Terminal (macOS/Linux) or PowerShell (Windows). Replace the model name if you picked a different one from the table.

ollama pull qwen3.5:4b

The first download can take several minutes depending on your connection. When it finishes, optionally load the model once to confirm:

ollama run qwen3.5:4b

You may see a chat prompt. Type /bye to exit. The model stays installed for The Novelist to use.

Keep Ollama running while you write. Quitting Ollama stops AI features until you launch it again.

Verify Ollama is working

1. Confirm the Ollama menu bar / tray icon is visible.
2. List installed models:

ollama list

You should see your model (e.g. qwen3.5:4b) in the output.

3. Optional API check: curl http://localhost:11434/api/tags — should list your model in JSON.

Connect Ollama in The Novelist

1. Enable the Local AI plugin

1. Open The Novelist and your workspace.
2. In the left sidebar, open Plugins.
3. Find Local AI (Ollama) and turn it on.

This unlocks AI drafting and rewrite tools in the editor. Without this plugin, Settings → AI Provider stays hidden.

2. Configure Settings → AI Provider

1. Open Settings (gear icon or sidebar).
2. Go to the AI Provider section.
3. Select Ollama (Local).
4. Base URL: leave as http://localhost:11434 unless you changed Ollama's port on purpose.
5. Model: enter the exact name you pulled, e.g. qwen3.5:4b (not llama3 unless you pulled that model).

3. Test and save

1. Click Test connection.
2. If the test succeeds, click Save.
3. Open a chapter in the editor — use the AI assistant or rewrite actions. They now call your local model.

Troubleshooting

“Connection failed” or test never succeeds

• Is Ollama running? Check the menu bar / system tray icon.
• Is the model name exact? Run ollama list and copy the name from the NAME column.
• Did you pull the model? Run ollama pull qwen3.5:4b again if unsure.
• On Linux, is the service up? systemctl --user status ollama
• On Windows, try visiting http://localhost:11434 in a browser — you should see "Ollama is running".

AI is very slow

• Try a smaller model (e.g. qwen3.5:0.8b or qwen3.5:4b).
• Close other heavy apps to free RAM.
• On Mac, consider an MLX variant from the model's page on ollama.com.

Settings → AI Provider is missing

• Enable Local AI (Ollama) under Plugins first.

ollama command not found

• Restart Terminal/PowerShell after install.
• On Windows, reboot once if the installer asked you to.
• Re-run the Linux install script or add Ollama to your PATH manually.

Related guides

• How to connect DeepSeek in The Novelist — cloud API instead of local Ollama.
• All guides
• Install The Novelist