Nod by n8n FAQ

Short answers, all drawn from Nod's official product page, documentation, and repository. Where an answer rests on n8n's own testing, we say so.

Do I need an n8n account to use Nod?

No. Nod is a standalone app, separate from n8n and from the agents you can build inside n8n. You don't register for n8n, and you don't need to set up a server or cloud service first — install it, connect your accounts, and start asking.

Is Nod free?

The software is. Nod is free to download and open source under the Apache License 2.0, with no subscription. What you pay for is the AI model it runs on: your provider bills you directly for API use, unless you sign in with a ChatGPT Plus or Pro plan you already have, or run a local model at no per-use cost. See Nod pricing for the full picture.

Could Nod run up a big model bill?

It has guardrails, and you set them. Every task shows its actual cost and finishes with a receipt; you can set a monthly spending limit that stops API-model work when it's reached; and if a single task's cost passes $5 mid-task (a threshold you can change), Nod stops and confirms with you first. Schedules show an estimated monthly cost before you enable them. One honest note from n8n's launch post: early testers found some tasks cost more than expected, and n8n says recent releases have brought costs down significantly.

Where does my data live?

On the computer or server running Nod: its database, files, memories, and saved logins stay there. Two qualifications worth knowing. First, app connections route through a connection service hosted by n8n and Composio by default. Second, new users share basic usage data with n8n by default — you can turn that off under Settings → Privacy & data (conversation traces are only sent if you separately switch them on). Your passwords sit in an encrypted Vault on your machine and are entered through private forms the model can't read.

Which AI models can I use?

Most of them. You add a connection under Settings → Models and pick. n8n lists 30+ providers, including OpenAI/ChatGPT, Anthropic, Google Gemini (AI Studio API key), OpenRouter, Google Vertex AI, Amazon Bedrock, Azure OpenAI, Cloudflare, and any OpenAI-compatible endpoint — plus names like DeepSeek, Groq, Mistral, and xAI. You can connect several providers, or even two accounts with the same provider, and use different models for chat, schedules, and background work. See Connect a model.

Can I use my ChatGPT Plus or Pro plan instead of an API key?

Yes — ChatGPT Plus and Pro plans can sign in as a model source, no separate API purchase needed. Two exceptions from the docs: Claude must be connected by API key (Anthropic's authentication conditions require it), and a Gemini app subscription can't be reused as a Nod sign-in (Google doesn't allow it), so Gemini needs an AI Studio API key.

Can I run Nod with a local model?

Yes, through Ollama, LM Studio, vLLM, or llama.cpp, connected as an OpenAI-compatible endpoint. Two hard requirements: the model must support tool calling, and it needs at least 64,000 tokens of context — Nod's instructions and tools take up roughly 20,000 tokens before your first message. Local-model tasks show as "No price" in the usage log and don't count against your monthly limit. One caveat from the docs: with a local model, the safety check may fall back to a small hosted model or decision service, which can still incur a small API cost — and n8n notes small models are less reliable at that check.

Does Nod work on Windows?

Yes, in preview. Nod supports Windows 11; its "private computer" there runs as a WSL distribution, app data lives under %APPDATA%\nod\profiles\local, and saved secrets are encrypted with Windows DPAPI. The Windows install guide covers setup, including the update quirks of early versions.

What about Mac and Linux?

Mac is the most established path: macOS 14 or later, on Apple silicon or Intel — see Install Nod on Mac. Linux x64 is available as a Flatpak preview; it needs Flatpak, a desktop secret store (GNOME Keyring or KWallet), and hardware virtualization with /dev/kvm access — and note that on Linux, Nod's built-in update check is off, so new versions are a manual reinstall.

Does Nod work offline?

Plan on being online. Nod itself runs locally and can drive a local model, but the jobs it does live in your online accounts and on websites, and app connections go through n8n's hosted connection service by default. Nothing in the official materials presents offline operation as a supported mode.

Will Nod ever act without asking me first?

For anything outside what you asked for, it's designed not to. Actions like sending, paying, posting, or deleting that aren't covered by your request and rules go through an independent check, and if they aren't covered, Nod asks you first. Web pages and emails can't grant it permission or give it orders. In n8n's own testing, it reports zero unpermitted actions across 678 runs — the maker's test result, not a guarantee, and n8n is upfront that the check is a model's judgment and can be wrong. That's why the asking exists.

Does Nod keep working when my laptop sleeps or is closed?

No. On a laptop, Nod stops when the machine sleeps, and everything — chat, channels, schedules — pauses while the app is closed. For always-on work, n8n's suggested setups are an always-awake machine (they name a Mac mini) or self-hosting on your own server. A managed hosting option is on the roadmap but isn't out yet.

What is the difference between Nod and n8n?

n8n is a workflow automation platform: you build automations, and it runs shared processes for a team or business. Nod is a separate personal agent for one person's errands — it is not the n8n AI Assistant, and using it doesn't change anything about n8n. If you already have n8n, you can optionally connect your instance so Nod can call your workflows. The Nod vs n8n page compares them properly.

Can Nod make phone calls?

Not yet. Making phone calls is listed by n8n as coming next — anything that needs a real voice call is still a job for you.

Can I share Nod with my team?

Not today. Each installation has a single owner, and Nod is designed for one person's accounts. A shared "Nod for your team" is on n8n's roadmap. If several people need automation now, that's the n8n platform's territory — see Nod vs n8n.

Still deciding? What is Nod explains the concept in full, and the Nod hub links every guide in the series.