Is n8n's Nod Free? Pricing, Setup, and Spending Limits Explained
n8n's Nod personal agent is free software, but you pay your model provider directly. Here's how costs, caps, and access actually work.
What is Nod and why does its pricing work differently?
Nod is a personal AI agent from n8n Labs, a spin-off of the n8n automation company. The software itself costs nothing to download or run. What you pay for is the model that powers it, because Nod lets you bring your own model provider instead of bundling one in behind a subscription fee. That single design choice separates it from most competing personal agents, which charge a flat monthly fee that already includes model access. With Nod, the app is free, but usage isn’t, and the bill goes straight to whichever provider you choose.
TL;DR
- Nod is free to download and run on your own Mac, Windows machine, or a personal server, with no license fee from n8n Labs.
- You pay your model provider directly, whether that’s Anthropic via API key, ChatGPT through a Plus or Pro login, OpenRouter, or a model running locally on your own hardware.
- Spending limits are built in, letting you set a monthly cap so costs for the model and any paid actions (like Apify scrapers) don’t run away from you.
- Every paid action requires explicit approval, including card payments where you confirm the security code each time and Nod never stores it.
- Regional availability depends on where it runs, since a self-hosted setup worked in the UK during testing, while rivals like Dots exclude the UK, Switzerland, and the EEA, and Claude Duet remains a US-only research preview.
- It connects to roughly 225 apps out of the box through Composio, and multiple accounts per app (several email inboxes, for example) are supported without extra cost from Nod itself.
- It’s open source, so the pricing model is transparent all the way down to the code, and nothing is hidden behind a paywall beyond the model usage itself.
How much does Nod actually cost to use?
There’s no license fee. The cost comes entirely from the model you connect and any third-party tools it calls on your behalf. During setup, Nod asks which model you want to run on: Claude through an Anthropic API key, ChatGPT through an existing Plus or Pro subscription, any model available through OpenRouter, or a model hosted on your own machine. Each of these has its own pricing structure set by the provider, not by n8n Labs. If you already pay for ChatGPT Plus, you can sign in and use that subscription rather than paying twice. If you use a local model, your only cost is the electricity and hardware you already own.
Beyond the model itself, some tasks cost money because they call paid services. In testing, finding and running an Apify scraper to collect lead data triggered a real charge on the Apify account connected through MCP, and the agent paused to get explicit approval before spending that money. That pattern, stopping before anything costs you, is consistent across both model usage and connected tools.
How does the spending limit work?
Nod shows you what each task costs as it runs the model and connected services, and you can set a monthly spending limit inside the app to cap total spend. This matters because an agent with access to your accounts, your card, and the ability to run paid scrapers or book services could rack up costs quickly without a ceiling in place. The spending limit acts as a hard stop rather than a soft warning, giving you a way to bound your exposure even if you’re running multiple scheduled tasks across several connected inboxes or accounts.
Payments themselves need a second layer of approval. Every card transaction requires you to enter your card security code at the time of payment, and Nod does not save that code afterward. Combined with the monthly cap, this creates two separate checkpoints before any money moves: a budget ceiling and a manual confirmation for each individual charge.
Is Nod worth it compared to paid personal agents?
The honest answer depends on how much you already spend on model access and how comfortable you are configuring your own stack. If you already have a ChatGPT Plus or Pro subscription, or an Anthropic API key you use elsewhere, Nod adds no extra subscription cost on top, just usage. Compare that to competitors like Dots, Muse, or Claude Duet, which typically bundle model access into a single paid plan. Nod’s model is more like paying for electricity instead of a fixed rent: cheaper if you’re a light user, potentially more expensive if you run it hard without watching the spending cap.
Nothing gets lost.
The bigger trade-off isn’t price, it’s control. Running on your own computer (or your own server, which Nod also supports) means your data and your agent’s “workspace” stay under your ownership rather than living on a vendor’s cloud infrastructure. That matches a broader pattern in how n8n has built its automation business: data ownership and transparency over convenience. The trust principles published alongside Nod, and the fact that the project is open source, reinforce that same positioning. You can read the code, audit how decisions get made, and see exactly what data goes where. That transparency isn’t something you get with most closed, subscription-based agents.
Where is Nod available, and does region affect pricing?
Nod’s availability depends on where you choose to run it rather than a region lock from n8n Labs. Because the desktop app runs on your own machine or your own server, a setup running locally worked in the UK during testing. That stands in contrast to Dots’ Pro plan, which is unavailable in the UK, Switzerland, and the EEA, and Claude Duet, which remains a US-only research preview with a waitlist. Muse is also still rolling out primarily in the US. Grok’s agent (Grokbot) was available in the UK and tested directly alongside Nod, producing comparable results on the same task.
This matters for pricing indirectly: if a competitor isn’t available in your region at all, the question of cost becomes moot. Nod’s self-hosted model sidesteps that problem for many users simply by not depending on a centralized cloud service tied to a specific country’s infrastructure.
What else factors into the real cost of running Nod?
Connecting your accounts is free. Nod uses Composio in the backend to offer one-click OAuth connections to around 225 apps, and unlike some competitors, it supports multiple accounts of the same app, such as several email inboxes, without additional charges. For anyone handling accounting or admin tasks that span multiple inboxes, that flexibility avoids needing a separate paid tier just to add a second account.
The main ongoing cost center, beyond model tokens, is any paid API or tool the agent uses to complete a task, like the Apify scraper example. Since Nod surfaces and confirms those costs before spending, it’s possible to keep usage close to zero if you stick to free connected apps and a lower-cost model, or scale up spend deliberately when a task genuinely needs it.
Frequently Asked Questions
Is Nod completely free to use?
The Nod software is free to download and run. You still pay for whichever AI model you connect it to (Anthropic, ChatGPT, OpenRouter, or a local model) and for any paid tools or services the agent uses during a task.
Can I use Nod without an API key?
Yes. You can sign in with an existing ChatGPT Plus or Pro subscription instead of using a separate API key, or run a model locally on your own computer at no per-use cost.
How do I stop Nod from overspending?
You set a monthly spending limit inside the app, and Nod shows the cost of each task before or as it runs. Payments also require manual card security code entry each time, so nothing charges automatically without your confirmation.
Does Nod work outside the US?
It depends on your setup rather than a fixed region restriction, since it runs on your own computer or server. It was tested successfully in the UK, a region where several competing personal agents are currently unavailable.
Is Nod open source?
Nothing gets lost.
Yes. The code is open source, which means you can inspect exactly how it handles payments, memory, and connected accounts, and modify it yourself if needed.
