n8n Labs Launches Nod: A Personal AI Agent That Lives on Your Laptop and Does Your Chores
n8n Labs has released Nod, a free open-source personal AI agent that installs on your Mac or Windows PC and autonomously handles repetitive tasks — portal logins, invoice downloads, price monitoring — using a sandboxed local browser. Indian professionals should note it requires an external AI provider API key, has no native Tally or RazorPay integrations, and stops working when your laptop sleeps.
A New Kind of AI Agent Runs Directly on Your Computer
Most AI tools give you a list of steps. Nod goes away and does them.
According to the n8n Labs announcement on blog.n8n.io, the n8n team has released Nod as a research preview — a standalone desktop application for Mac and Windows that acts as a personal AI agent. It is separate from n8n the workflow platform, requires no n8n account, and installs on your machine like any ordinary app. You tell it what you need done, and it works inside your apps and on websites — in a sandboxed virtual browser running locally — until the job is finished.
The n8n team describes Nod as being for the “single player” — the individual, the solo operator, the freelancer — in contrast to n8n itself, which is built for shared team processes. It is open source under the Apache 2.0 licence, and the code is publicly available on GitHub.
What Nod Actually Does, in Plain Language
Think of Nod as a junior assistant who can operate a computer on your behalf. It has its own sandboxed browser and local virtual environment. You connect it to Gmail, Google Calendar, Google Drive, and Slack, and it can also navigate websites that have no API or integration at all — the kind of supplier portal or government web form that no automation tool has ever supported cleanly.
The n8n team describes several tasks Nod can handle:
- Cancelling a subscription, including turning down retention offers and requesting a refund
- Watching a price on a website and alerting you when it drops
- Finding a hotel that matches your trip criteria and booking it once you confirm
- Logging into supplier portals each month, downloading invoices, and filing them for your accountant
- Pulling numbers from a dashboard that offers no export function and dropping them into your weekly report
Nod also works proactively: it reads your connected apps and suggests tasks it could take off your plate. You hold a single conversation with it — through the app itself, or through Telegram or WhatsApp — while it runs multiple jobs in parallel in the background. Repeating tasks, like a morning briefing or a daily price check, can be scheduled.
On memory, the n8n team notes that Nod scored 93.2% on OnLongMemEval, a public benchmark for long-term memory in AI systems, placing it among the best published systems on that test. It learns your communication style from Gmail so its email drafts sound like you, and it records how it completed each task so it can repeat the process more reliably next time.
A Concrete Scenario: The Ahmedabad Export Consultant
Consider a self-employed export documentation consultant in Ahmedabad who handles paperwork for ten to fifteen small textile clients every month. Each client has accounts on two or three different freight and customs portals — none of which talk to each other, none of which offer a bulk export function.
Currently, the consultant logs into each portal individually, downloads the month’s documents, renames them according to client convention, and drops them into a shared Google Drive folder. It takes the better part of a day each month.
With Nod installed on a Windows desktop, she could describe this workflow once in plain language: which portals, what login credentials to use, how to name files, and where to put them. According to the n8n Labs announcement, Nod encrypts saved logins on the local machine and enters them through private forms the underlying AI model cannot read. The agent would then handle those portal logins each month on a schedule, download the invoices, and file them — without exposing passwords to any external service.
If a portal changes its layout mid-year — which freight portals do — Nod records a note about how the site changed and adjusts its approach. The consultant only steps in when Nod flags that something is outside the instructions she originally gave.
This is exactly the category of work — repetitive, login-dependent, portal-heavy, requiring no judgment — that Nod is designed for.
What It Costs, and the Model-Provider Catch
Nod itself is free and open source. You pay nothing to download and run it. What you do pay is your AI provider — OpenAI, Anthropic, or whichever API you connect. The n8n team acknowledges in the announcement that early testers found task costs higher than expected, and notes that recent releases have brought those costs down substantially.
To help you stay in control, Nod includes guardrails: every task displays its cost, a completed task ends with a receipt, you can set a monthly spending cap, and Nod checks in with you once a task crosses USD 5 (approximately ₹425 at current rates). Each scheduled task also shows an estimated monthly cost before you activate it.
For Indian users, this means two things. First, billing is in US dollars through your AI provider, with no INR invoicing or GST-compliant billing from Nod itself. Second, you must be comfortable generating and managing API keys from providers like OpenAI or Anthropic — or running an open-source model locally. The n8n team is explicit that Nod currently suits people comfortable with this kind of setup.
Limitations You Should Know Before You Start
What to Watch For
Nod is a research preview, and the n8n team is explicit that it is an experiment from their Labs division, separate from the core product roadmap. The honest framing is that this is software you should approach with curiosity and reasonable caution — not production-critical infrastructure.
If you are a solo professional whose biggest time drain is logging into portals, downloading files, filling forms, or monitoring prices on websites, Nod is worth exploring. Download it for Mac or Windows from nod.n8n.io, connect it to a low-stakes account first — perhaps a test Gmail inbox — and hand it one repeating task. Watch what it does, check the cost receipt it generates, and decide from there.
Keep an eye on the n8n community forum’s Labs category, where the team is actively collecting feedback. The hosted version — which would remove the “laptop must stay awake” problem — is the feature most likely to change how useful Nod becomes for Indian users who cannot dedicate a machine to running it overnight.
