How much does it cost to keep n8n on a VPS?

Minimum configuration: VPS for 1 vCPU + 1 GB RAM - 500-700 rubles/month. This is enough for 5-10 simultaneous workflows. n8n.cloud - $20/month, if you don’t want to administer the server. AI API - another 5-10 $/month for typical tasks.

Is n8n free or do you have to pay?

Self-hosted n8n is completely free - no restrictions on nodes or executions. There is a paid n8n.cloud and an enterprise version, but 90% of marketing scenarios are completed with free self-hosted VPS.

Which AI is better to connect to n8n for marketing?

For Russian content - Claude Sonnet 4.6. For analytics and tables - GPT-4o. For a budget option - Claude Haiku or GPT-4o-mini.

Is N8n difficult to set up without programming experience?

The entire interface is drag-and-drop. Nodes are configured through a form, not through code. The only technical point is deployment to VPS via Docker, but there is only one command.

Which Yandex APIs can be connected to n8n?

Ya.Metrica is a ready-made connector node. Yandex Direct - via HTTP Request node with OAuth token. VK Ads - via VK API. TG Bot, Google Sheets, Notion, Slack - all connect in 2 clicks.

How reliable is self-hosted n8n for production?

When configured correctly, the n8n works for months without failure. On my projects it crashed 2 times in six months - both times due to the disk being full of logs. Solve logrotate in 5 minutes.

What to do if n8n workflow did not work?

The built-in execution log shows at which step the error occurred. Most often: incorrect API key, AI token limit exceeded, input JSON structure changed.

Are there ready-made templates for marketing in n8n?

The built-in n8n library contains 50+ marketing templates. Plus, the community publishes ready-made configs on GitHub. Below in the article are 10 scripts for import.

All articles

AIMay 26, 202610 min

10 n8n scenarios for AI marketing automation 2026: from leads to content

10 ready-made workflows n8n for AI marketing automation: auto-generation of content, CPL monitoring with alerts in Telegram, auto-reports from Ya.Metrika, TG funnel for lead generation, competitor parsing. Self-hosted on VPS for 500 rub/month.

N8n is the most useful thing in the 2026 marketer’s AI stack, which few people know about. It connects AI, Telegram, Yandex Metrica, Google Sheets and Notion into one pipeline without a line of code. I have collected 10 scenarios that actually work on my projects - with configs and savings metrics.

I've been working with n8n since the summer of 2024. The current 6 projects run 15 workflows: from auto-generating posts to monitoring CPL across channels. The numbers in the article are the median savings of hours on my streams.

N8n is not just another no-code constructor. This is the only tool where the AI agent, Telegram bot and Yandex API live in the same pipeline without spacers. If your workflow requires more than 15 minutes of manual launch, you are already paying for what n8n does for free.

1. Automatic generation of content from the price list via Claude

You take Google Sheets with products, n8n runs each line through Claude, AI generates a post for the TG channel, a description for the card, a title for the ad. The output is a finished table with 3 options for each text. Previously: copywriter 6 hours for 30 products. Now: n8n does it in 8 minutes.

On EZ KATKA, this script generates promotion descriptions for 19 arenas - 26 texts every week instead of manual writing.

2. CPL monitoring with alerts in Telegram

Workflow pulls the Yandex Direct and VK Ads APIs once an hour, calculates the CPL for campaigns, and compares them with the threshold value. If the CPL is higher than the norm, an alert is sent to Telegram. If it’s lower, there’s silence, don’t pull.

On a project with a budget of 1.5 million rubles/month, this scenario saved about 120 thousand in the first month - the CPL jump was caught on the day of the change, and not a week later.

3. Auto-reports from Ya.Metrica and Yandex Direct in Telegram

A weekly dashboard that fills itself out: expense, leads, CPL, CPC, conversion - for each project. N8n connects Ya.Metrika, Yandex Direct and Google Sheets. The result is a page in Notion that I can log into on Monday morning and see the numbers without manual copying.

4. Lead generation funnel in TG bot

The user writes to the bot, n8n receives the message via a webhook, passes it through Claude - AI determines the intent of the request. Next, there is a different scenario for each intent: purchase → qualification, price → link to price list, complaint → escalation per person.

I discussed this scenario in detail in an article about Telegram bot for marketing. The conversion from message to manager contact increased from 8% to 27%.

5. Lead scoring via Claude AI from CRM

N8n checks new deals in amoCRM every 30 minutes, uploads lead data, and runs it through Claude. The AI returns a score of 1-10 and the reason for the score. Before: Managers spent 20-30 minutes a day sorting. After: Touch conversion on golden leads increased by 35%.

The same approach - in AI agents in n8n, where I discussed lead scoring in more detail.

6. Competitor parsing and price monitoring

Workflow goes through 10-20 competitors’ pages once a day, parses HTML, runs it through AI, which highlights changes. Before: manual walkthrough every 2 weeks, missed 70% of changes. After: alert on the day of change.

7. UTM generator with duplicate checking

The team fills out Google Forms, n8n generates a UTM link, checks for duplicates, writes to Google Sheets. Before: every fourth link has an error. After: 0 errors. About standards - UTM tags 2026.

8. Sentiment monitoring of comments and reviews

Workflow collects comments from the TG channel, reviews from Yandex.Maps, mentions from vc.ru and Habr, and runs it through AI. If there is a negative with a tonality higher than 7/10 - an instant alert. On one project, the manager responded to a negative review within an hour, and the review was edited to 4 stars.

9. AI generation of a content plan from a brand guide

You upload the brand guide, Tone of Voice, the last 20 posts and the price list to Google Docs. N8n sends everything to Claude, AI generates 30 posts per month. Before: content plan 4-6 hours. After: draft in 20 minutes.

10. Autonomous AI agent for routine issues

N8n + Claude Agent work as a virtual assistant: they receive a question from TG, AI determines what data is needed, pulls the appropriate APIs, and generates an answer. Not just a knowledge base bot - an agent that collects data from 3-4 sources.

Scenario	Save time	API cost/month
1. Automatic content generation	5-6 hours/week	5-8 $
2. CPL monitoring	3-4 hours/week	0 $
3. Auto-reports	2-3 hours/week	0 $
4. TG lead funnel	10-15 hours/week	8-12 $
5. Leadscoring from CRM	2-3 hours/week	5-8 $
6. Parsing competitors	2-4 hours/week	2-3 $
7. UTM generator	1-2 hours/week	0 $
8. Sentiment monitoring	3-5 hours/week	5-8 $
9. AI content plan	4-6 hours/month	3-5 $
10. AI agent question and answer	2-3 hours/week	10-15 $

11. Conclusion

AI automation in n8n is not about technology, but about time. Each workflow above takes 30 minutes to set up and saves 2-10 hours per week. On 6 projects, 15 workflows save me about 40 hours a month.

You don't have to run all 10 at once. Take one scenario - CPL monitoring or auto-report - set it up in an hour and see if it works.

If you want to analyze your specific situation, write to Telegram or through form. Starting consultation - 0 rub.