Host game servers and AI model endpoints on the hardware you already own.
One small agent turns the machine you already own into a managed platform — game servers for Minecraft, Rust, Palworld and more, or private AI model endpoints powered by Ollama. No port forwarding, no cloud bills, no GPU required for AI.
# Run in PowerShell — installs & pairs the agent
irm https://install.nodemesh.app/install.ps1 | iex
✓ Agent installed & connected
# Run in your terminal — installs & pairs the agent
curl -sSL https://install.nodemesh.app | sudo sh
✓ Agent installed & connected
# Same agent — AI runtime auto-configured
curl -sSL https://install.nodemesh.app | sudo sh
✓ Agent installed & connected
✓ Ollama runtime detected
✓ Endpoint ready at api.nodemesh.app/e/…
🔒 No shell runs on your Linux host. Downloads are checksum-verified. → How we keep your hardware safe
// What you can run
Install the NodeMesh agent on any machine. Choose what to run — game servers, AI model endpoints, or both.
// Game Servers
Host Minecraft, Rust, Palworld, Valheim and more on your own PC or server. NodeMesh handles the control plane — traffic proxy, auto-restart, console, file manager, backups. Your players connect; your machine does the work.
// AI Endpoints
Run open-source LLMs on your own machine with an OpenAI-compatible REST API. No GPU required. Your data never leaves your hardware — no per-token billing, no cloud inference, no request logging.
// Features
Your hardware does the work. NodeMesh handles the hard parts.
Install the agent on any machine — gaming PC, spare laptop, home server. Your hardware runs the games; we run the control plane.
No port forwarding, no exposed IP. Player traffic routes through our edge to your machine, so your home network stays private and your real address never leaves the panel.
Minecraft, Valheim, Hytale, Rust, Palworld, Satisfactory, Project Zomboid and ARK: Survival Evolved — runtimes auto-installed, server downloaded, configured for you. See the full game list.
Servers restart automatically on crash and the agent reconnects after network drops — so your world stays up while you're away.
Your players connect to a memorable name like yourworld.mc.nodemesh.app — no IP, no port to remember. Minecraft connects with no port at all; your machine's real IP stays hidden.
Live console streaming, an in-browser file manager, mods, and a visual config editor. Manage everything without ever touching SSH.
// AI Endpoints
NewYour prompts and completions never leave your hardware. NodeMesh deploys the model to your own node — no cloud inference, no request logging, no third-party data pipeline.
Every endpoint speaks the OpenAI chat completions protocol. Point any SDK, LangChain integration, or HTTP client at your node URL — nothing else changes. POST /v1/chat/completions
No per-token billing, no monthly API quota, no usage cap. Once the model runs on your node, every request is free. Build chatbots, generate embeddings at scale — the bill doesn't move.
// Try it
# OpenAI-compatible — works with any client
curl https://api.nodemesh.app/e/{id}/v1/chat/completions \
-H "Authorization: Bearer nm_..." \
-d '{"model":"llama3.2:3b","messages":[{"role":"user","content":"Hello"}]}'
→ {"choices":[{"message":{"content":"Hi! How can I help?"}}]}
# Swap model name — same endpoint, same API
curl https://api.nodemesh.app/e/{id}/v1/chat/completions \
-H "Authorization: Bearer nm_..." \
-d '{"model":"qwen2.5:7b","messages":[{"role":"user","content":"Explain gRPC"}]}'
→ {"choices":[{"message":{"content":"gRPC is a high-performance RPC framework…"}}]}
# Embeddings endpoint — same node, different path
curl https://api.nodemesh.app/e/{id}/v1/embeddings \
-H "Authorization: Bearer nm_..." \
-d '{"model":"nomic-embed-text","input":"NodeMesh game servers"}'
→ {"data":[{"embedding":[0.021,-0.043,…]}]}
// Setup
Three steps. No networking knowledge required.
Paste one line from your dashboard. It downloads, installs as a service, and pairs automatically.
Pick a game and player count. NodeMesh installs the runtime and the server for you — watch it live in the console.
Your connection address appears in the panel. Send it to friends — no IP, no port forwarding to explain.
// AI Endpoints
Same agent, same install. Pick a model and get an API URL.
The same one-line install works for both game servers and AI endpoints. The agent auto-detects your hardware and configures the Ollama runtime.
Choose from Llama 3.2 3B, Llama 3.1 8B, Qwen 2.5 7B, or Nomic Embed. NodeMesh pulls the model to your node and starts serving it automatically.
Your endpoint URL appears in the dashboard. Use it with any OpenAI SDK, LangChain, or plain HTTP. Private, unlimited, no billing.
// Supported games
New games ship as templates — they appear here automatically when they land.
// Live now
survival
survival
sandbox
sandbox
survival
survival
survival
building
survival
More games coming.
// Already using something else?
Coming from a free host or a bare tunnel? Here's the honest side-by-side.
// Beta pricing
All features included — no credit card, no time limits, no catch.
BYOH — Free
Free · 1st node
Extra nodes: 5 NMC/month (NodeCoin credits, granted during beta — no payment needed).
Your hardware. Our control plane. Game servers and AI endpoints — on the machine you already own.
Get started — it's free