Quickstart

Install PawFlow and reach the first chat.

Use the Docker installer for a first run. It prepares persistent data, starts the bootstrap server, and opens the browser wizard for the admin account and first LLM service.

Docker install Source install

PawFlow installer preview — Replace this with `assets/media/product/quickstart-installer.webp` when the final wizard image is ready.

Recommended

Release installer

The fastest path is the latest release installer zip. Download it, unzip it, then let the install script pull the matching server and relay images. Current release: .

terminal

Loading current release command...

What it does

Uses the release installer package, pulls the matching prebuilt PawFlow and relay images, prepares runtime files, and creates persistent data under `~/pawflow`.

Where to open

After the server starts, open `https://localhost:19990/install` for the first-run wizard.

Release downloads

All downloadable packages live on the current GitHub release. The website builds these links from the single release version configured in `site.js`.

Installer zip PawCode Linux zip PawCode Windows zip PawCode Debian package Relay CLI Linux zip Relay CLI Windows zip Relay Desktop AppImage Relay Desktop Windows VS Code extension VSIX All release assetsOpen the GitHub release page for checksums, tarballs, source archives, and platform variants.

Complete the installer wizard

Accept the local certificate

The bootstrap server starts with a self-signed certificate for private/local installs.

Enter the bootstrap key

The temporary first-run Private Gateway key is `RoyBetty`. The wizard forces you to replace it before finalizing.

Create the runtime

Create the admin user, configure the first LLM service and summarizer service, deploy the PawFlow Agent flow, and open the starter conversation with `assistant` selected.

Connect relays and media services

After the server is running, the next useful setup is usually a relay boundary and optional media/voice services.

Install a server relayCreate a managed relay service from PawFlow for server-side tools and diagnostics. Install a remote relayUse Relay Desktop or Relay CLI from the release downloads on the machine that owns the files or desktop. Access the desktopOpen the Desktop Relay view, including local desktop access when the relay allows local host tools. Add media servicesConfigure image, video, audio, TTS, STT, voice clone, and speech-to-video providers.

Source or development install

Use source mode when you are developing PawFlow itself or want to inspect/edit the Python runtime directly.

terminal

git clone https://github.com/allcolor/PawFlow-Agents.git
cd PawFlow-Agents
pip install -r requirements.txt
python cli.py start --host 0.0.0.0 --port PORT

Next steps

Install a server relayCreate the managed server-side relay service. Install a remote relayConnect Relay Desktop or Relay CLI to a workstation. Open desktop accessGive agents controlled access to a full desktop. Start the first agentOpen a conversation and select the right LLM service. Add media/voiceEnable image, video, audio, TTS, STT, and voice services. Check securityUnderstand relays, providers, secrets, and permissions.