How to use PlayOn
Install Home, add nodes if you need capacity, wire a model or MCP agent, start a game.
Quick start
Install PlayOn, create the Owner, pick a provider, add nodes if you need capacity, share /play.
Getting started with AI
Local Ollama or a cloud preset — min and recommended models so in-app chat can install a server.
Nodes
Add a LAN spare or any cloud VPS — PlayOn handles WireGuard so guests still join on the LAN.
LLM providers
Venice, NVIDIA, OpenAI, Ollama, and the other Settings presets — get a key and paste it in.
External agents (MCP)
Connect Cursor, Claude Code, Codex, OpenClaw, or Hermes to PlayOn over MCP.
All guides
- Quick start — Install PlayOn, pick a provider, add capacity if you need it, start a game, share the player link.
- Getting started with AI — Pick local Ollama or a cloud preset so the in-app agent can create, start, and stop a game server.
- Nodes (LAN + cloud) — Add spare PCs or any VPS as compute — PlayOn configures WireGuard for cloud so players still join on the LAN.
- Docker on nodes — Why Docker matters for containerized games, how PlayOn installs it on Linux, and what to do on Windows.
- For players — RSVP on the LAN hub, open the player panel, and join the game.
- Packages and the catalog — What a package is, how to install games from playon.games, and how agents use them.
- Changelog — Notable PlayOn Home releases and how in-app updates work.
- LLM providers — Choose a cloud or local model for in-app agents. Keys stay on your host.
- Venice — Privacy-focused OpenAI-compatible API — common default for PlayOn in-app agents.
- NVIDIA — Sign up for NVIDIA’s API catalog, create a key, and use it as your PlayOn LLM provider.
- OpenAI — Use an OpenAI API key for in-app PlayOn agents.
- Anthropic — Use Claude via Anthropic’s OpenAI-compatible endpoint in PlayOn Settings.
- Google Gemini — Use a Google AI Studio API key with Gemini’s OpenAI-compatible endpoint.
- OpenRouter — Route many models through one OpenRouter API key in PlayOn.
- DeepSeek — Use DeepSeek’s API as the in-app agent provider.
- Ollama (offline) — Run a local model when the LAN has no cloud API access.
- Custom OpenAI-compatible — Point PlayOn at any OpenAI-compatible base URL and model.
- External agents (MCP) — Let Cursor, Claude, Codex, OpenClaw, Hermes, or any MCP client set up and manage your game servers.
- Cursor — Connect Cursor to PlayOn over MCP so the agent can bootstrap and manage game servers.
- Claude Code — Point Claude Code at PlayOn MCP to install games and manage servers from the terminal.
- Codex — Connect OpenAI Codex (or Codex CLI) to PlayOn MCP for LAN bootstrap and ops.
- OpenClaw — Point OpenClaw at PlayOn so your agent can bootstrap and manage game servers.
- Hermes Agent — Connect Hermes Agent to PlayOn MCP to bootstrap and operate game servers.
- Any MCP client — Generic Streamable HTTP setup for PlayOn’s /mcp endpoint.