---
title: 'n8n Nod is INSANE! (FREE!) 🤯'
source: 'https://youtube.com/watch?v=metLPrUNv5g'
video_id: 'metLPrUNv5g'
date: 2026-11-17
duration_sec: 719
channel: 'Julian Goldie SEO'
---

# n8n Nod is INSANE! (FREE!) 🤯

> Source: [n8n Nod is INSANE! (FREE!) 🤯](https://youtube.com/watch?v=metLPrUNv5g)

## Summary

The video introduces n8n Nod, a free, open-source personal AI agent that operates on your own computer with a private sandbox, browser, and workspace. It demonstrates five practical use cases—from sales call preparation to subscription management—and highlights key safety features like requiring user approval before any action and logging all tasks.

### Key Points

- **Introduction to n8n Nod** [00:03] — Nod is a free, open-source personal AI agent that runs on Mac or Windows, preps sales calls, chases follow-ups, and finds forgotten tool trials. It asks before sending anything, including Slack replies.
- **Launch and Basics** [00:30] — Nod launched as a research preview on November 10th from n8n Labs. It's a standalone app (no workflow building or n8n account needed), free to download and use, with code on GitHub under Apache 2.0. You bring your own AI account (e.g., ChatGPT plan or API key from OpenAI, Anthropic, or OpenRouter). Version 0.8.2 landed on November 16th.
- **Unique Feature: Own Computer** [01:28] — Nod sets up a private sandbox on your machine—a sealed-off mini-computer with its own browser, workspace, and files. It can interact with websites that have no official API, clicking through them like a human. For apps like Gmail, Calendar, Drive, Slack, and Notion, you connect once via sign-in; passwords are typed through a private form the AI never reads.
- **Safety: Asks Before Acting** [02:22] — Nod has a built-in check that stops before sending, posting, deleting, or buying anything you didn't clearly ask for—even a Slack reply waits for your yes. Every task leaves a record of sign-ins, app actions, and steps taken, so you can verify completion.
- **Performance Stats** [02:51] — n8n tested Nod on test apps and websites across 226 errands (each run three times), achieving 98% completion with zero unsafe actions. These are n8n's own numbers on simulated accounts; the safety check is an AI model, so manual verification is advised initially.
- **Old vs. New Workday** [04:01] — The old way: opening multiple tabs, writing follow-ups from scratch, forgetting quotes, updating CRM from memory. The new way: Nod reads inbox/calendar/website to provide one-page briefs, drafts follow-ups in your voice, spots unanswered quotes, and you just approve or edit.
- **Use Case 1: Software Founder** [04:45] — Nod creates morning briefs for demo calls by reading email threads and visiting prospect websites, then drafts follow-ups after calls, learning your writing style from Gmail. Every draft waits for your approval.
- **Use Case 2: Outside Sales Rep** [05:42] — Via WhatsApp, you can ask which quotes haven't had replies in 5 days; Nod checks sent mail, drafts nudges, and logs visits in CRM (even without API by signing into the website). It also preps tomorrow's visits from your calendar.
- **Use Case 3: Coach/Consultant** [06:40] — Nod finds forgotten tool trials and subscriptions in Gmail, can cancel them (including declining 'please stay' offers), and handles session rescheduling with reminders via Telegram or WhatsApp.
- **Use Case 4: Creator** [07:23] — Nod researches topics by reading web pages and PDFs, writing findings to a document. You can set weekly scheduled jobs (e.g., every Monday at 8) to check websites and report only changes. Since version 0.8.0, you can assign a lighter AI model to scheduled jobs to save usage limits. Catch: schedules only run while the computer is awake.
- **Use Case 5: Real Estate** [08:20] — Nod checks for new inquiries, drafts replies with property answers and viewing times, and remembers rules (e.g., 'Anna handles weekend viewings'). You can edit or delete memory notes.
- **Addressing Doubts** [09:03] — Safety: Nod asks first, reads can't give new orders, everything is logged; start with drafts/read-only jobs. Too many tools: Nod consolidates, closing tabs; give it a week. Not technical: no workflow building, ~10-minute setup; limits include one owner per install, no phone calls, stops when laptop sleeps.
- **Action Steps** [10:03] — 1) Download Nod from nod.n8.io, connect AI account. 2) Connect one app (Gmail or calendar). 3) Give one small errand (e.g., write a brief). 4) Turn best errand into a scheduled job with a lighter model. Start small, one errand at a time.

### Conclusion

n8n Nod is a promising free AI agent that automates repetitive tasks across your apps and websites, with strong safety features. Its 98% success rate in tests and practical use cases make it worth trying, but users should start small and verify outputs.

## Transcript

personal AI agent that gets its own computer. [music] It's free, it's open- source, and it runs on your Mac or Windows PC. It preps your sales calls, [music] chases your follow-ups, and finds the tool trials you forgot to
cancel. And before it sends anything, even a Slack reply, [music] it asks you first. By the end of this video, you'll know five ways to put it to work [music] and the one setting that stops your AI limits running dry. Let's get into it.
Let me start with what actually happened. On the 10th of November, NAT launched NOD as a research preview. It comes from NAT Labs, a brand new team inside the automation company built to try ideas outside the main product. So,
NOD is its own app. You don't build a workflow and you don't need an NAT account. You download it, connect your apps, and ask for something in plain English. The code is public on GitHub under the Apache 2.0 license. So, the
app itself is free to download and free to use. The one thing you bring is your to use. The one thing you bring is your own AI account like a chat GPT plan or a key from a provider such as OpenAI, Anthropic, or Open Routter. Version
0.8.2 already landed on the 16th of November. Research preview means it's early and it can make mistakes, so you start small, and I'll show you how. Here's what makes Nod different from a normal chatbot. It
gets its own computer. When you install it, Nod sets up a private sandbox on your machine. A sandbox is a sealed off little computer that lives inside yours.
It has its own browser, its own workspace, and its own files. So, when you ask it to pull numbers from a dashboard with no export button, it opens that site in its own browser and clicks through it the way you would.
That works even on websites with no official connection at all. For the apps that do, like Gmail, Google Calendar, Drive, Slack, and Notion, you connect them once with a sign-in. Your files, your memory, and your saved login stay
on your machine, and when it needs a password, it types it in through a private form that the AI model never gets to read. Now, the thing a lot of you keep asking for inside the boardroom. Agents that ask before they
run anything, NOD has that builtin. Before it sends, posts, deletes, or buys anything you didn't clearly ask for, a separate check stops and asks you. Even a Slack reply waits for your yes. Picture every errand as a little lantern
sitting on the ground, and it only lifts into the sky when you nod. And here's how you know it really did the work. Every task leaves a record of the sign-ins, app actions, and steps it took. So when it says done, you can see
took. So when it says done, you can see the proof. N8 tested this themselves on the proof. N8 tested this themselves on test apps and test websites. Across 226 errands, each one run three times. NOD finished 98% of them with zero unsafe
actions. Those are NAT's own numbers on simulated accounts. And the safety check is itself an AI model, so still check its first few jobs yourself. If you want NOD doing real errands in your business without weeks of trial and error, that's
exactly what we work on inside the AI profit boardroom. You get the NOD errand playbook with the exact prompts for morning call briefs, follow-up drafts, and quote chasing. You get the automation that chains them together, so
your briefs, your follow-ups, and your scheduled jobs run as one system, and every send still waits for your yes. Weekly coaching calls let you bring your exact errand and set it up live. Daily tutorials show you each new NOD update
tutorials show you each new NOD update the week it ships alongside 3,400 plus stuff. And if you want to learn how to make videos like this one, the whole faceless video system we use is in there too. The link is in the comments and
description or go to aiprofitboardroom.com. Here's how this changes a normal working day. The old way looks like this. You open three tabs before every call and write the same follow-up from scratch
again and again. Open quotes sit in your scent folder until you remember them a week later. You update the CRM at night from memory and a tool trial you forgot about quietly rolls over. Now, the new way. It reads the inbox, the calendar,
and the website and hands you a one-page brief before each call. It drafts the follow-up in your voice and waits. It spots the quote nobody answered and drafts the nudge. And your part is saying yes or saying change this line
and the clicking happens without you. Let's start with a software founder. The pain SAS builders keep posting in the boardroom is this. The demos get booked and then the leads go cold. Here's what Nod does about it. You ask it for a
morning brief for every demo call on your calendar today. It opens your calendar, reads the email thread with each prospect, visits their website in its own browser, and writes a one-page brief for each call. Who they are, what
they asked about, and what to bring up. NAT shows this exact job on their own product page. Then, after the calls, you message it. Draft the follow-ups for today's demos. If you've connected Gmail, it learns your writing style from
the emails you've already sent, so the drafts sound like you. Every draft waits for you to tweak a line and say yes. Why it matters? You walk into every demo prepared and every demo gets a follow-up the same day, which is exactly where a
lot of deals slip away. Next, an outside sales rep. A member asked a simple question this week. What should I actually do next? Picture your week. Quotes sent, driving between visits, admin piling up. You text nod on
WhatsApp from the car park. which quotes haven't had a reply in 5 days. It checks your sent mail, finds them, and drafts a short, friendly nudge for each one. They all wait for your yes. Then you ask it to log today's visits in your CRM. If
your CRM has no app connection, NOD signs into the website in its own browser and fills it in there with your saved login typed through that private form. At the end of the day, prep tomorrow's visits, and it pulls a short
list from your calendar of who, where, and what to bring up. Why it matters. The follow-ups go out, the CRM stays up to date, and you start tomorrow knowing to date, and you start tomorrow knowing exactly what to do next. Now, a coach or
a consultant. This one's for anyone with shiny object overwhelm. You've signed up for a lot of AI tools this year, and some are trials you've forgotten about. Ask Nod. find every tool trial and subscription email in my inbox from the
last 3 months and tell me which ones I still use. It searches your Gmail, gives you a list, and can cancel the ones you don't need, including turning down the please stay offer, which is one of NAT's own example jobs. For your sessions,
message it on Telegram or WhatsApp. Move Thursday's session with Priya to Friday at 10:00 and send her a reminder. It checks your calendar, drafts the change and the reminder and asks before anything goes out. Why it matters? Fewer
tools sitting in the background and your clients get their reminders without you opening five different apps. Next, a creator. And this is where scheduled jobs come in. Ask Nod to research a topic and it reads web pages and PDFs
and writes the findings up as a document in its workspace with the links. Then turn it into a weekly job. Every Monday at 8, check these five websites and tell me only what changed. It only reports back when something actually moves. And
here's the setting I promised you for anyone who keeps hitting usage limits. Since version 0.8.0, you can pick a different AI model for your scheduled jobs than for your everyday tasks. Put a lighter model on
the background jobs and save your main model for the work that needs it. One honest catch, schedules only run while the computer running nod is awake. Close the laptop and the job stops, which is why some people run it on an always on
Mac Mini. Why it matters? Your research is waiting on Monday morning and you never started it. And one more for real estate. A new inquiry comes in on a listing at 9 at night. Ask Nod to check for new inquiries and draft a reply to
each one. Answer their question about the property. Offer two viewing times from your calendar and keep it in your tone. The draft sits there waiting. You see it on your phone. You say yes and it goes. And because it remembers what you
tell it, you only explain your rules once. Say Anna handles weekend viewings and it remembers that from then on. And you can open its memory and edit or you can open its memory and edit or delete any note. Why it matters? Fast
personal replies. And you're still the one who decides what gets sent. Now, you might be sitting there with a few doubts. Doubt one, is it safe to let an agent into my accounts? Nod is built to ask first before it sends, posts,
deletes, or buys anything you didn't ask for. A web page or email it reads can't give it new orders, and everything it does is logged. Still, start with drafts and readonly jobs, and give it more once you've watched it work. Doubt two, I
already have too many tools, so why add another one? Nod is there to close tabs. One chat on your desktop or your phone working across the apps you already use. Give it one week and if it doesn't hand you back real time, uninstall it. Doubt
three. I'm not technical. You don't build a workflow and NAT says setup takes about 10 minutes. Install the app, connect your accounts, and ask. Just know the limits. One owner per install, no phone calls yet, and it stops when
the laptop sleeps. So, here's what you can do this week. Step one, download NOD from nod.n8.io for Mac or Windows and connect your AI
account. Step two, connect one app, just Gmail or just your calendar. Step three, give it one small errand, like write me a brief for tomorrow's first call, and read what comes back. Step four, turn your best errand into a scheduled job
and put a lighter model on it. one errand at a time. And if you want help doing that, join the AI profit boardroom. You'll get the NOD errand playbook with the exact prompts for morning call briefs, follow-up drafts,
quote chasing, trial cleanups, and inquiry replies, plus the automation that chains them into one daily routine. From the morning brief to tomorrow's prep with every send waiting on your nod. Bring your own errand to a coaching
call and we'll set it up with you. You'll find 3,400 plus members, a 30-day NOD road map, and daily tutorials as NOD keeps shipping updates, and the faceless video system behind this video is in there, too, so you can learn to make
The link is in the comments and description, or go to a profitboardroom.com. So, that's N8N, NOD, a free open-source personal agent from N8N Labs. It gets its own computer and browser right on
your Mac or Windows PC. It does errands in your apps and on websites, and it asks before it sends. It runs scheduled jobs. It remembers what you tell it, and you can message it on Telegram or WhatsApp. And in NAT's own tests, it
WhatsApp. And in NAT's own tests, it finished 98% of 226 errands with zero unsafe actions. Drop a comment and tell me, would you handn your follow-ups first or your morning call briefs? I'll see you in the next one.
