Install Phantomline desktop

The full studio that runs locally on your machine. Uses your own Ollama, Kokoro, and Forge for unlimited high-quality renders. License-key activated. Auto-starts on boot so phantomline.xyz can switch to your local install with one click.

Easiest path: download & double-clickyour OS detected

Phantomline desktop runs as a tiny background service on your computer (auto-starts on boot, lives in your system tray). When you visit phantomline.xyz from any browser, the 'Local Phantomline' button in the studio toggle opens your local install in a new tab. Projects sync between the two via your account so the same library follows you everywhere.

Download Phantomline source (Python · Windows · macOS · Linux) ~30 MB

Windows PowerShell. Paste & run.

Invoke-WebRequest https://phantomline.xyz/download/phantomline-source.zip -OutFile phantomline.zip; Expand-Archive phantomline.zip -DestinationPath . -Force; cd phantomline; py -3.11 -m venv .venv; .\.venv\Scripts\pip install -r requirements.txt; .\.venv\Scripts\python server.py

Open Windows Terminal or PowerShell, paste the line above, hit Enter. The script handles Python check, install, and model download.

macOS Linux Terminal. Paste & run.

curl -L https://phantomline.xyz/download/phantomline-source.zip -o phantomline.zip && unzip phantomline.zip && cd phantomline && python3 -m venv .venv && .venv/bin/pip install -r requirements.txt && .venv/bin/python server.py

Open Terminal, paste, hit Enter.

Paste this into Claude Code (works in Cursor / Codex / Cline too)

Open Claude (the desktop app or web), click + New session, switch to the Code tab, then paste the prompt below as your first message. Claude does the install, downloads, and verification for you. No terminal commands you have to type yourself.

Get Claude (free) Claude Code quickstart

Install Phantomline desktop on this machine so I can use my local Ollama for high-quality video generation. Set up auto-start so it always runs in the background and the phantomline.xyz toggle can switch to it with one click.

1. Verify Python 3.11+ is installed (run `python --version` or `py -3.11 --version` on Windows). If missing, install Python 3.11 from python.org or via Homebrew/apt. Phantomline does not work on Python 3.12+ yet (Forge dependency).
2. Download the Phantomline source ZIP from https://phantomline.xyz/download/phantomline-source.zip into a permanent location (NOT a temp dir — auto-start will reference this path forever). Use curl/wget on macOS/Linux: `curl -L https://phantomline.xyz/download/phantomline-source.zip -o phantomline.zip`. Or Invoke-WebRequest on Windows: `Invoke-WebRequest https://phantomline.xyz/download/phantomline-source.zip -OutFile phantomline.zip`. The ZIP is self-hosted on phantomline.xyz so no GitHub account is needed.
3. Extract the ZIP. macOS/Linux: `unzip phantomline.zip`. Windows: `Expand-Archive phantomline.zip -DestinationPath . -Force`. This creates a `phantomline/` folder.
4. cd into the `phantomline` directory.
5. Create a virtual environment: `python -m venv .venv` (use `py -3.11 -m venv .venv` on Windows). Activate it: `source .venv/bin/activate` on macOS/Linux, `.venv\Scripts\activate` on Windows.
6. Install Python dependencies: `pip install -r requirements.txt`. This takes 5-10 min.
7. Confirm Ollama is running: `ollama list` should show llama3.1 (or pull it with `ollama pull llama3.1`).
8. SET UP AUTO-START so server.py runs every time the user logs in. Use the right method for the OS:
   - Windows: Create a Scheduled Task that runs at user logon. Use `schtasks /create /tn Phantomline /tr "<absolute path to .venv\Scripts\pythonw.exe>\" \"<absolute path to server.py>\"" /sc onlogon /rl highest`. Use pythonw.exe (not python.exe) so no terminal window appears.
   - macOS: Create ~/Library/LaunchAgents/com.phantomline.server.plist with <plist><dict><key>Label</key><string>com.phantomline.server</string><key>ProgramArguments</key><array><string>/absolute/path/to/.venv/bin/python</string><string>/absolute/path/to/server.py</string></array><key>RunAtLoad</key><true/><key>KeepAlive</key><true/></dict></plist>. Then run `launchctl load ~/Library/LaunchAgents/com.phantomline.server.plist`.
   - Linux: Create ~/.config/systemd/user/phantomline.service with [Unit]\nDescription=Phantomline\n[Service]\nExecStart=/absolute/path/to/.venv/bin/python /absolute/path/to/server.py\nRestart=always\n[Install]\nWantedBy=default.target. Then run `systemctl --user enable --now phantomline`.
9. Verify auto-start works: open http://localhost:5000/api/health in a browser. Should return JSON with `"ok": true`. If yes, the background service is up.
10. Open http://localhost:5000/account, sign in with my Google account. License auto-issues as Free tier (5 renders/month) or activates whatever paid license is associated with my email.
11. Tell me when it's done so I can open phantomline.xyz, click the studio header toggle, and switch to my local install with one click.

What Claude will do for you:

Detect your OS
Check whether prerequisites are installed (Python, Git, etc.)
Install anything missing
Install Phantomline desktop and download model weights
Verify the install worked
Tell you when it's done

You'll see Claude ask for permission before running each install command. You can approve all or step through.

1Install Python 3.11

Phantomline desktop runs on Python. Get Python 3.11 from python.org/downloads (NOT 3.12+ yet, has Forge compatibility issues). No Git or GitHub account needed — we host the download directly.

python --version

2Download Phantomline (ZIP, self-hosted)

Direct download from phantomline.xyz. No GitHub login required. Pick a folder with ~5 GB free (Phantomline + Python deps). Or just click the big download button at the top of this page.

# macOS / Linux:
curl -L https://phantomline.xyz/download/phantomline-source.zip -o phantomline.zip
unzip phantomline.zip
cd phantomline

# Windows (PowerShell):
Invoke-WebRequest https://phantomline.xyz/download/phantomline-source.zip -OutFile phantomline.zip
Expand-Archive phantomline.zip -DestinationPath . -Force
cd phantomline

3Install Phantomline's Python dependencies

Use a virtual environment so Phantomline's libs don't collide with your system Python. Takes 5-10 min on first install.

python -m venv .venv
source .venv/bin/activate   # macOS/Linux
# .venv\Scripts\activate      # Windows
pip install -r requirements.txt

4Install Ollama if you haven't yet

Phantomline uses your local Ollama for script + idea + title generation. Get the friendly install guide at /install/ollama.

ollama pull llama3.1

5Start Phantomline

Easiest: double-click the launcher file in your phantomline folder.

Windows: double-click start-phantomline.bat
macOS: double-click start-phantomline.command (right-click → Open the first time)
Linux: run ./start-phantomline.sh

The launcher activates the venv, starts the server on port 5000, and opens your browser automatically. Leave the terminal window open — closing it stops Phantomline. (For an always-on background service, see step 6.)

# Or, if you'd rather run it manually:
python server.py

6Set up auto-start so it always runs in the background

Once-and-done. After this, Phantomline runs on every login and the phantomline.xyz toggle can switch to your local install with one click. Pick the command for your OS. Replace <ABSOLUTE_PATH> with the full path to your phantomline folder.

# Windows (PowerShell as Administrator):
schtasks /create /tn Phantomline /tr "<ABSOLUTE_PATH>\.venv\Scripts\pythonw.exe <ABSOLUTE_PATH>\server.py" /sc onlogon /rl highest

# macOS (creates a LaunchAgent):
cat > ~/Library/LaunchAgents/com.phantomline.server.plist <<'EOF'
<?xml version="1.0" encoding="UTF-8"?>
<plist version="1.0"><dict>
  <key>Label</key><string>com.phantomline.server</string>
  <key>ProgramArguments</key>
  <array>
    <string><ABSOLUTE_PATH>/.venv/bin/python</string>
    <string><ABSOLUTE_PATH>/server.py</string>
  </array>
  <key>RunAtLoad</key><true/>
  <key>KeepAlive</key><true/>
</dict></plist>
EOF
launchctl load ~/Library/LaunchAgents/com.phantomline.server.plist

# Linux (systemd user service):
mkdir -p ~/.config/systemd/user
cat > ~/.config/systemd/user/phantomline.service <<EOF
[Unit]
Description=Phantomline desktop server
[Service]
ExecStart=<ABSOLUTE_PATH>/.venv/bin/python <ABSOLUTE_PATH>/server.py
Restart=always
[Install]
WantedBy=default.target
EOF
systemctl --user enable --now phantomline

7Sign in with your license

Open http://localhost:5000/account, sign in with the same Google account. Free tier (5 renders/month) auto-issues; paid licenses activate from your account email. Your projects + library will sync to phantomline.xyz so you can view them from any device.

Verify it worked

Open http://localhost:5000/app in your browser. You should see the Phantomline studio with your local Ollama detected. Make a quick test video to confirm everything works.

curl http://localhost:5000/api/health

Then: Reload /app — the readiness checklist will flip Phantomline desktop to READY.

Stuck? Email support@phantomline.xyz with your OS and the error and we'll send you exact steps.