# Node Lodge

> NVIDIA GPU machines on demand, billed by the minute. Node Lodge Secure runs in partner datacenters, Community on independent providers' machines, each session in its own virtual machine.

Each page below is plain Markdown. The same pages, as HTML, are at https://nodelodge.com/docs.

## Get started

- [Quickstart](https://nodelodge.com/docs/quickstart.md): From access request to a shell on a GPU: create your account, add funds, pick a machine in the catalog, launch it, connect over SSH, stop it.
- [Pick a GPU for your model](https://nodelodge.com/docs/pick-a-gpu.md): Describe your workload, get the machines it fits, cheapest first. How the GPU memory is computed, what the result covers and where it stops.
- [SSH keys](https://nodelodge.com/docs/ssh-keys.md): Create an SSH key on macOS, Linux or Windows, copy its public half into the console, and fix the permission errors that block a connection.
- [Connect over SSH](https://nodelodge.com/docs/connect-ssh.md): Create a key, allow your IP, connect as ubuntu. JupyterLab through SSH on Secure, in the browser on Community. Tunnels and connection errors.

## Run workloads

- [Sessions and limits](https://nodelodge.com/docs/sessions-and-limits.md): How a session runs: a budget held on prepaid credit, Budget Guard, automatic end, deletion of the machine, interruptible capacity, account limits.
- [GPU sessions from a notebook](https://nodelodge.com/docs/notebooks.md): Launch a GPU machine from Jupyter with a maximum cost, get the SSH and JupyterLab links, and have it stopped when you are done.

## Secure and Community

- [Secure and Community](https://nodelodge.com/docs/secure-vs-community.md): Two tiers, one dedicated VM per session. Where each runs, who controls the hardware, and which one fits your work.

## Billing

- [Billing](https://nodelodge.com/docs/billing.md): Prepaid credit that never expires, a budget on every session, Budget Guard, real-time charges, automatic top-up, refunds, disputes and taxes.

## Developers

- [API quickstart](https://nodelodge.com/docs/api.md): Create an API key, authenticate with a Bearer token, check it with whoami, choose scopes, find every endpoint, and revoke a key you no longer use.
- [API reference](https://nodelodge.com/docs/api-reference.md): Every public endpoint of the Node Lodge API: authentication, parameters, request bodies and a curl example for each.
- [CLI and SDK](https://nodelodge.com/docs/cli-and-sdk.md): Rent GPU machines from a terminal with the nodelodge command, or from your own code with the TypeScript and Python SDKs.
- [Use Node Lodge from an AI agent (MCP)](https://nodelodge.com/docs/mcp.md): Let Claude Code, Claude Desktop or your own agent find a GPU, launch a session under a hard spending cap, connect over SSH and stop it.
- [GPU tools for agent frameworks](https://nodelodge.com/docs/agent-frameworks.md): Give a LangChain, LlamaIndex, OpenAI Agents SDK or Claude Agent SDK agent GPU tools with hard cost caps, and stop its sessions on exit.
- [Run GPU jobs from GitHub Actions](https://nodelodge.com/docs/github-action.md): Use the nodelodge/run action to run a command on a GPU from a workflow, under a hard cost cap, and always stop the machine.
- [Errors and rate limits](https://nodelodge.com/docs/errors-and-rate-limits.md): Every error code the Node Lodge API returns, what each one means and what to do, and the request limits per minute and per hour.

## Providers

- [Provider requirements](https://nodelodge.com/docs/provider-requirements.md): What a Community machine needs, how to set up the firmware (virtualization, IOMMU) and VFIO on Ubuntu, and how to check it in a few commands.
- [Pair a machine](https://nodelodge.com/docs/pair-a-machine.md): Link a machine to your provider account with the Node Lodge agent on Linux or Windows: get it, check it, pair it, keep it reporting.
- [Providers](https://nodelodge.com/docs/providers.md): List your NVIDIA machine on Community: requirements, pairing with the agent, what the review checks, the internet speed test, pricing and your 85 percent share.

## Account and security

- [Account limits](https://nodelodge.com/docs/account-limits.md): How many GPUs your account runs at once, which GPU classes it can start, how much credit it adds per day, and how the limits rise.
- [Notifications](https://nodelodge.com/docs/notifications.md): What Node Lodge e-mails you, what stays in the console, and how to change it in Notification settings.

## Troubleshooting

- [SSH connection refused](https://nodelodge.com/docs/troubleshooting-ssh.md): Fix an SSH connection that times out, is refused or denied: the IP rule of a Secure machine, the key, the user, and a changed host key.
- [Session does not start](https://nodelodge.com/docs/troubleshooting-session-start.md): A launch refused, or a session stuck on starting or marked failed: what each message means, what you are charged, and what to do next.
- [Budget reached](https://nodelodge.com/docs/troubleshooting-budget.md): Why a session stopped before its budget, what you were charged, how to raise the budget of a running session, and how to get warned before the end.
- [Payment declined](https://nodelodge.com/docs/troubleshooting-payment.md): A refused credit purchase, automatic top-up or card charge: what stops, what keeps running, and how to reopen the account.
- [FAQ](https://nodelodge.com/docs/faq.md): Access, billing, Secure and Community, what happens to your data, interruptible machines, SSH access, providers and acceptable use.

## Optional

- [Full documentation in one file](https://nodelodge.com/llms-full.txt)
- [OpenAPI 3.1 description of the API](https://nodelodge.com/v1/openapi.json)