What Is n8n? Is It Free, How to Install It, Cloud Pricing, License, AI and Turkish Support

In n8n you connect apps on a visual canvas and build workflows that automate a repetitive task; it also has features for building AI agents. There are two ways to use it: n8n Cloud (hosted by n8n, 14-day free trial) or on your own server (the Community edition is free; the server, security and updates are on you). The source code is available, but it is not an OSI-approved open source license; it uses the “Sustainable Use License”, and commercial use is limited. The official sources don't mention a Turkish interface; there is no Turkish lira (TRY) pricing.

What it does: Connects apps on a visual canvas and automates tasks
Free: Yes on your own server (Community) · Cloud 14-day trial
Turkish interface: Not stated in the official sources
Where: Browser (Cloud) or your own server (Docker)
Current version: 2.43.5 stable (October 14, 2026)
License: Fair-code, not OSI-approved open source
Price: Cloud Starter €20/month (billed yearly) · no TRY price

What is n8n for? Can I use it without code?

Can I use it without code?

Yes, but by the right route. n8n Cloud needs no installation: you sign up in the browser and build the workflow on a drag-and-drop canvas. The official docs recommend Cloud for people without technical knowledge. Installing on your own server, on the other hand, requires server, container (Docker) and security know-how; the docs say a faulty setup can lead to data loss, security issues and downtime.

You don't have to write code, but you can add JavaScript or Python code to a workflow if you want (“Code steps in workflows” in the pricing table).

n8n (short for “nodemation”, pronounced “n-eight-n”) makes another task happen automatically when something occurs: a new form response arrives, add a row to a spreadsheet, send a message to the team. This flow is called a workflow; each step is a node, and the first node that starts the flow is called the trigger.

TriggerSchedule, webhook, form, new email…

NodesAdd a record, send a message, transform data, run code…

ResultAn app gets updated or someone gets notified.

n8n counts one full run of a workflow from start to finish as one “execution”; the number of steps doesn't matter.

Connecting apps and automatingWhat you get: A workflow that runs on a triggerBuilt into the app Running it free on your own serverWhat you get: n8n running at localhost:5678Needs a separate setup Building an AI agentWhat you get: An agent that uses toolsBuilt into the app

Who it's for: businesses, agencies and developers who repeat the same tasks by hand and would prefer to keep their data on their own server. According to the README, there are more than 1500 integrations and more than 9,000 ready-made workflow templates.

Source: n8n README (github.com/n8n-io/n8n), docs.n8n.io “Choose how to use n8n”; October 14, 2026.

How do you install n8n? Cloud, Docker, npm and your own server

There are four routes. If you don't code, start with Cloud; if you want to try it on your own computer, the Docker route is the most current. The commands below are taken verbatim from the official docs and README.

Sign up. The quickstart in the docs gives app.n8n.cloud/register. The pricing page says “No credit card required” for the trial.
Start the trial. The 14-day trial gives you the Pro plan's features, but with a limit of 1,000 executions and 5 concurrent executions.
Open your first workflow. Start with Start from Scratch or Create Workflow; add a trigger node as the first step.

If you don't upgrade when the trial ends, your workspace is deleted; you can download your workflows within 90 days after the trial ends.

First install Docker (Docker Desktop is available for Mac, Windows and Linux). On Windows, the docs recommend using WSL for the Docker Compose setup. Then run the one-line setup command:

One-line setup (docs: One-line setup)› curl -fsSL https://get.n8n.io | sh

This command checks that Docker is running, creates an n8n folder in your current directory, writes the config files, downloads and starts n8n, and tells you the address: http://localhost:5678. It comes with an SQLite database and helper services for the Assistant (you turn the Assistant on later). If you want to read the script before running it, the docs also explain how to download and inspect it.

If you want a more manual setup, the Docker command from the README:

With Docker (README)

› docker volume create n8n_data
docker run -it --rm --name n8n -p 5678:5678 -v n8n_data:/home/node/.n8n docker.n8n.io/n8nio/n8n

Updating a one-line setup› curl -fsSL https://get.n8n.io | sh -s -- --upgrade

The Docker page is marked “outdated” in the docs and points to the Docker Compose page; for a new installation, the one-line setup or Compose is recommended. The database is SQLite by default; for serious use, the docs recommend PostgreSQL.

Only for a quick test on your own computer. Requires Node.js; the docs give the version as 20.19 to 24.x.

Try without installing› npx n8n

Global install and start

› npm install n8n -g
n8n start

Warnings in the docs: npm installation is deprecated as of n8n 3.0 (new installs are pointed to the one-line setup), it isn't safe for production, and n8n Assistant doesn't work on this route.

The docs provide separate guides for installing on cloud providers: DigitalOcean, Heroku, Hetzner, AWS, Azure, Google Cloud Run, Google Kubernetes and OpenShift; plus Docker Compose. Ready-made n8n templates from providers such as Hostinger are not listed in the n8n docs; for those, see the provider's own page.

Time zone: the GENERIC_TIMEZONE setting matters for schedule nodes; the default is America/New_York.
Database: PostgreSQL is recommended over SQLite in production (the docs explain the configuration on the Compose page).
Interface language: self-hosted n8n has an N8N_DEFAULT_LOCALE setting, but the docs don't list which languages are available (see the Turkish section).

The docs' own warning: self-hosting is for experts; mistakes can lead to data loss, security issues and downtime. If you lack experience, Cloud is recommended.

Current version and updates

As of October 14, 2026, the latest stable release on GitHub is 2.43.5 and the beta is 2.44.1. n8n ships a new minor release almost every week; the stable release is for production, the beta is the newest but can be unstable. On your own server you do the updates yourself; on Cloud you choose version and update settings in the dashboard.

Source: docs.n8n.io (One-line setup, Install with Docker, Install using Docker Compose, Install with npm, Update your version), n8n README, GitHub releases; October 14, 2026.

Server and hardware for n8n: how much RAM do you need?

What the official docs say

The Docker Compose setup guide asks for at least 4 GB RAM and 2 vCPU; this is for the stack that includes the Assistant's code-running sandbox service. For bare n8n, the docs don't give minimum hardware.

Cloud's own limits

Starter 320 MiB RAM, Pro 640 MiB (10K-execution tier) or 1,280 MiB (50K), Enterprise 4,096 MiB; 100 GB storage per instance. The docs say the n8n software alone uses 180 MiB RAM on average.

On Windows

For the Compose setup on Windows, the docs recommend using WSL (Docker Desktop with WSL2, or Docker inside Linux).

Real cost

The software is free; server cost, backups, security and updates are on you. The official sources don't give a server cost; calculate it from your provider's pricing page.

Source: docs.n8n.io “Install using Docker Compose”, “Manage your data”; October 14, 2026.

How do you use n8n? First workflow and commonly requested connections

The official quickstart has you build this workflow: every Monday morning, pull NASA's solar flare data and send it down one of two paths depending on the flare class. The point isn't the task itself; it's to teach the concepts of trigger, credential, condition and expression.

Open a new workflow. Start from Scratch or Create Workflow.
Add a trigger. Add first step > Schedule Trigger; set the interval to weekly, Monday 09:00.
Add an app node and create a credential. In the example it's the NASA node; with Create new credential you paste the key you got from the app. A credential is defined once in n8n.
Add a condition. Split into two paths based on the data with the If node (in the example, if the “classType” field contains X).
Send the result somewhere. In the example it's the PostBin node; in your own work it would be a Gmail, Slack or spreadsheet node.
Test it. Run a single step with Execute step and the whole flow with Execute Workflow; see the output in the side panel.
Publish. After you click Publish, the workflow runs automatically according to its trigger.

New form responseTrigger: form or webhook

Write to a spreadsheetGoogle Sheets node

Notify the teamSlack or Gmail node

This is an example flow. Search the integrations list in the n8n docs for which operations (nodes) each app has.

Commonly requested connections (with nodes in the docs)

WhatsApp: the docs have a “WhatsApp Business Cloud” node and a “WhatsApp Trigger”. A WhatsApp Business account and Meta-side setup are also required; details are on the node's docs page.
Telegram: there are “Telegram” and “Telegram Trigger” nodes. Triggers that use webhooks require n8n to be reachable from the internet (see the Common problems section).
Instagram: no separate Instagram node appears in the docs; the Facebook Trigger node's docs include an “Instagram” object. Check the limits of what it can do on that page.
Gmail, Slack, Google Sheets: the docs have nodes for them.

You can also start from a ready-made template (n8n.io/workflows). The pricing table shows templates and all integrations (nodes) as “included” on every plan.

Having n8n Assistant build a workflow (example)"Every morning at 9, find the new rows added to the “Customers” sheet in Google Sheets and send the name and email from each row to the #sales channel in Slack."n8n Assistant is in preview and can make mistakes; check the workflow it builds before publishing it. The official sources don't say how it behaves when you write in Turkish.

Source: docs.n8n.io “Build your first workflow”, “Use n8n Assistant”, integration node pages; October 14, 2026.

n8n AI features: AI Agent, Assistant and MCP

n8n itself is not an AI model; you connect the model and tools to the workflow. The README says it can connect to OpenAI, Anthropic, Google or open models, and that you can switch providers without changing the architecture. Parts that are in preview are marked.

AI Agent node

You connect a chat model and at least one tool; the agent decides on its own which tool to call to complete the task. In the pricing table, “Agents, MCP and AI nodes” appears from Starter up.

n8n Assistant (preview)

It builds, edits, tests and debugs workflows through chat; the result is a normal workflow. It is available on Starter and Pro in Cloud, and on Community, Registered Community and Business when self-hosted; since it's in preview, it can make mistakes. When self-hosting, you add the model provider's key yourself. Monthly credits: Starter 1,600, Pro 4,000 or 9,600.

AI Workflow Builder

You describe your goal in natural language; it handles choosing, placing and configuring the nodes. Each request counts as one credit; failed or stopped requests don't count.

MCP server (Claude, ChatGPT and coding agents)

With n8n's built-in MCP server enabled, a tool like Claude or ChatGPT can connect to an n8n instance and build workflows, edit them (editing existing workflows from n8n 2.13 on), manage data tables and run them. Example configurations in the docs: Lovable, Claude Desktop, Claude Code, Codex, Gemini CLI, Cursor, VS Code, Windsurf.

MCP Server Trigger and MCP Client

There are separate nodes for turning a single workflow into its own MCP server (MCP Server Trigger) or connecting an agent to an external MCP server (the registry in the nodes panel or the MCP Client Tool).

n8nClaudeVia the built-in MCP serverChatGPTNamed as an example tool in the docsClaude CodeConfiguration example in the docsCursorConfiguration example in the docs

Get to know Claude: Claude guide

Get to know ChatGPT: ChatGPT guide

The docs don't separately state which plans include MCP access. Agents and the Assistant can make mistakes; check the workflow before publishing it, and don't paste secrets such as keys or passwords into the chat.

Source: docs.n8n.io (AI Agent, Use n8n Assistant, AI Workflow Builder, Use n8n MCP server, MCP servers), n8n README, n8n.io/pricing; October 14, 2026.

Is n8n free? Cloud plans, execution counting and euro pricing

The software itself is free on your own server (Community). Cloud is paid; there's only a 14-day trial. Prices are shown in euros and based on yearly billing; the pricing page says you save 17% with yearly billing. The monthly billing price is unclear in the official sources.

Community (your own server)Free

Almost all features
No execution counter
Free folders and debugging with registration (Registered)
SSO, LDAP
Environments, version control with Git
Projects and workflow sharing

Cloud Starter€20 /month · billed yearly · 2,500 executions

1 shared project, 5 concurrent executions
AI agents, MCP, AI nodes
AI models without your own key, 1,600 Assistant credits/month
Forum support
Admin roles
Insights dashboard

Cloud Pro€50 /month · billed yearly · 10,000 executions

Everything in Starter
3 shared projects, up to 50 concurrent
7 days of insights, admin roles
4,000 or 9,600 Assistant credits/month
SSO and LDAP
Multiple environments, Git version control

Business (self-hosted only)€667 /month · billed yearly · 40,000 executions

Everything in Pro
SSO and LDAP, 6 projects
Multiple environments, Git, queue mode
Hosting by n8n (not on Cloud yet)
Dedicated support

EnterpriseContact sales

Everything in Business
Unlimited projects, 200+ concurrent
Log streaming, external secrets store
Support agreement (SLA)

View as a table (Cloud limits)

	Starter	Pro	Enterprise
Executions/month (pricing page, starting tier)	2,500	10,000	Custom
Concurrent executions	5	20 / 50	200+
Max duration of one execution	5 min	40 min	40 min
Saved executions	2.5K	25K	50K
Execution log retention	7 days	30 days	Unlimited
Workflow history	1 day	5 days	365+ days
Shared projects	1	3	Unlimited

What counts as an execution?

n8n counts one full run of a workflow from start to finish as one execution; it doesn't matter how many nodes it contains or how much data it processes. To estimate usage, the pricing page says: a schedule that runs once a day is 30-31 executions a month, and a schedule that runs every 5 minutes is about 8,600-8,900 executions a month. For a webhook, the number of events per day is the number of executions per day.

Reaching the saved-executions limit doesn't stop workflows from running; it only affects the stored history. On Business and Enterprise, workflows keep running when the quota is exceeded, and extra charges may apply. The official docs don't say what happens on Cloud when the monthly execution quota runs out.

Free trial

14 days with the Pro plan's features; 1,000 executions, 5 concurrent executions, 180-second timeout, 1 team project, 800 Assistant credits.
No credit card required. If you don't upgrade, the trial ends on its own, you aren't charged and your workspace is deleted (you can download your workflows within 90 days).
Not included in the trial: admin roles, the project editor role, named versions, workflow diffs and the n8n public API.

Real cost on your own server

The software is free; server cost and updates are on you. The official sources don't say what you'll pay for a server; check the pricing page of the cloud or hosting provider you choose. If you register your email with “Community” (“Unlock” under Settings › "Usage and plan"), folders, debugging in the editor and custom execution data are unlocked for free.

Payment, tax, TRY and discounts

No Turkish lira (TRY) price: prices are in euros; there's no official Turkish page.
The payment processor is Paddle.com; n8n says it doesn't store payment method details. The pricing page doesn't say whether tax is included; check the total before paying. The official sources don't say whether card payments from Turkey go through.
Startup Plan: 50% off the Business plan for companies with fewer than 20 employees. The page also has an Education link; the terms are on that page.

Source: n8n.io/pricing (October 14, 2026); docs.n8n.io “Try free then choose a plan”, “Compare editions”, “Manage your data”. The data docs mention two Pro tiers of 10K and 50K executions; the pricing page showed only 10K.

Official pricing page last checked on October 28, 2026.

Is n8n open source? License and commercial use

n8n's source code is available on GitHub, but its license is the Sustainable Use License (a “fair-code” model). It is not an OSI-approved open source license; GitHub also shows the license as “Other”. According to the README, part of the code (files with “.ee.” in their names) additionally requires an n8n Enterprise License. The summary below is from LICENSE.md; don't make a commercial decision without reading the text itself. If you receive a violation notice and end the violation within 30 days, the license is reinstated.

Allowed

Using, copying and modifying the software for your own internal business purposes or for non-commercial and personal use
Distributing or providing it to others, but only free of charge and for non-commercial purposes

Not allowed

Removing or obscuring license, copyright and similar notices
Distributing or providing the software to others for a fee or for commercial purposes (LICENSE.md doesn't permit this)
Using source files with “.ee.” in their names without a valid n8n Enterprise License

If you're an agency or consultant, LICENSE.md doesn't say whether setting up n8n on behalf of a client counts as “your own internal business purposes”. If you plan to host n8n and sell it to clients as a service, write to the license address given in the README (license@n8n.io) to clarify. This is not legal advice.

Source: github.com/n8n-io/n8n LICENSE.md and README (October 14, 2026); docs.n8n.io “Compare editions”.

Does n8n work in Turkish?

Interface (menus)Unclear	The official sources don't mention a Turkish interface. Self-hosted n8n has an `N8N_DEFAULT_LOCALE` setting, but the supported languages aren't listed; as of October 14, 2026, the official GitHub repository had only the en.json language file. Check for yourself whether Cloud has a language setting.
DocumentationUnclear	The docs pages are in English; no official Turkish translation appears.
Turkish replies in a workflowUnclear	n8n doesn't generate replies; the AI model you connect does. Turkish quality depends on that model, and the n8n docs don't address it separately. Test with a small example first.
Price displayNo	Prices are in euros; no TRY price in the official sources.

Source: docs.n8n.io “Timezone and localization”; github.com/n8n-io/n8n (locale files); October 14, 2026.

What happens to the data I put into n8n? Data region and telemetry

Cloud: data in the EU

The pricing page says data on hosted plans is stored in the EU, on servers in Frankfurt, Germany. On your own server, the data stays wherever you choose. On Cloud, n8n is both data controller and processor; a data processing agreement (DPA) is part of the terms.

Telemetry

On your own server, telemetry is on by default; n8n says it doesn't collect data passing through workflows, credential contents or error payloads. On Cloud, telemetry can't be turned off, and sessions in the interface are recorded (deleted after 21 days; credential values aren't recorded).

Turning off telemetry on your own server

N8N_DIAGNOSTICS_ENABLED=false
N8N_VERSION_NOTIFICATIONS_ENABLED=false
N8N_TEMPLATES_ENABLED=false
With a paid license, usage counts continue to be sent with the license ID; this can't be turned off.

What is sent with AI features

The Assistant receives the workflow's nodes, data schema, node settings and expressions in code; credentials and the actual data processed are not sent. It is stated that it isn't used for model training.

Account deletion

When you delete the account from the Cloud dashboard, the workspace, credentials, user and execution data are deleted the same day and removed from backups within 90 days. Deletion requests go to help@n8n.io.

KVKK and self-hosting

There is no separate official statement for KVKK (Turkey's Personal Data Protection Law). When self-hosted, n8n is neither data controller nor processor; deleting user data is up to you. EXECUTIONS_DATA_MAX_AGE is recommended for deleting execution data automatically. This is not legal advice.

Source: docs.n8n.io “Privacy”, “Control telemetry”, “Use n8n Assistant”; n8n.io/pricing FAQ; October 14, 2026.

Common n8n problems and fixes

Symptom	What to do, per the official docs
A trigger from an external service like GitHub doesn't reach local n8n	Triggers that use webhooks require n8n to be reachable from the internet. For local development, the docs describe a cloudflared tunnel (requires Docker); this is for development only.
My workflows disappeared when I stopped Docker	The `-v n8n_data:/home/node/.n8n` volume in the command stores your data; if you remove it, the data is gone. PostgreSQL is recommended for serious use.
A scheduled flow runs at the wrong time	Set `GENERIC_TIMEZONE` and `TZ` to your own time zone; the default is `America/New_York`.
Executions are waiting on Cloud	There's a concurrent execution limit based on your plan (Starter 5); the rest are queued and processed in order.
The Cloud instance ran out of memory or storage	Split data into smaller chunks, don't save successful executions, break the workflow into sub-workflows, use fewer Code nodes. Execution history is deleted automatically based on your plan.
The database opened empty after upgrading Postgres to version 18	According to the docs, Postgres 18 writes data to a different location; don't delete the `PGDATA` line. Moving from an older version to 18 is a major version upgrade, so take a backup first.

Security advisories and updates

As of October 14, 2026, n8n's GitHub repository had 225 published security advisories; new medium-severity advisories were published that same day. This isn't a declaration of a defect but a process the project runs in the open; however, on your own server you do the updates yourself. Follow the security advisories and release notes and stay on the stable release.

Source: docs.n8n.io (Install with Docker, Install with npm, Install using Docker Compose, Understand concurrency, Manage your data, Timezone and localization); github.com/n8n-io/n8n/security/advisories; October 14, 2026.

How do you cancel an n8n subscription, and are there refunds?

Trial

You don't need to cancel the trial. It ends on its own when the period is over, you aren't charged and your workspace is deleted; you can download your workflows within 90 days.

Business plan

With a monthly commitment you can cancel any time; the subscription ends at the end of the current billing period. With a yearly commitment, cancellation takes effect at the end of the yearly term. n8n doesn't give partial (prorated) refunds. Invoice and bank transfer payment is available for yearly Business and Enterprise.

Cancellation steps and refund terms for Cloud Starter and Pro are not stated in the official docs. Before deciding, check the Upgrade and payment screen in your account and, if needed, n8n support.

Source: n8n.io/pricing FAQ and docs.n8n.io “Try free then choose a plan”; October 14, 2026.

n8n or Make?

Both connect apps on a visual canvas; the real difference is what they count and where they run. n8n counts one full run of a workflow from start to finish as one execution, and you can run the software on your own server if you want (Community, free; commercial use limited by the license). For Make's billing unit and plans, see the Make guide; because the units differ, the two products' numbers aren't in the same table.

Get to know Make: Make guide

The official sources don't say which is better. Work out your monthly task count in executions and test it in the Cloud trial. Source: n8n.io/pricing.

Frequently asked questions

What is n8n and what is it for?

n8n is a workflow and AI agent platform that connects apps on a visual canvas and automates repetitive tasks. When an event happens (for example, a form response arrives), another task is done automatically. These flows are called workflows.

Is n8n free?

If you install it on your own server, the software is free (Community edition); server cost, security and updates are on you. n8n Cloud is paid: there's a 14-day trial, then Starter is €20 a month and Pro €50 with yearly billing. There's no permanent free plan for Cloud.

How do I install n8n?

The easiest way, with Docker installed, is one line: curl -fsSL https://get.n8n.io | sh. Then open http://localhost:5678 in the browser. If you want no code and no installation, sign up for n8n Cloud. Installing with npm is deprecated as of n8n 3.0.

How do I use n8n?

You open a new workflow and first add a trigger node (such as a schedule, webhook or form), then connect app nodes and enter credentials. You test with Execute Workflow and go live with Publish.

How much does n8n Cloud cost?

On the pricing page with yearly billing, Starter is €20/month (2,500 executions) and Pro €50/month (10,000 executions); Business at €667/month (40,000 executions) is for self-hosting only. Prices are in euros; there's no TRY price. For the monthly billing price, select the monthly option on the pricing page.

Does n8n work in Turkish?

The official sources don't mention a Turkish interface; the docs are in English. Turkish replies in a workflow depend on the AI model you connect.

Can I build an AI agent with n8n and connect Claude or ChatGPT?

Yes: you build a tool-using agent with the AI Agent node. With the built-in MCP server, tools like Claude and ChatGPT can build and run workflows in n8n. The docs don't separately state which plans include MCP access.

Can I automate WhatsApp, Telegram or Instagram with n8n?

The docs have WhatsApp Business Cloud and Telegram nodes and triggers. There's no separate node for Instagram; an Instagram object is mentioned in the Facebook Trigger node. Account and permission requirements are on each node's docs page.

Is n8n open source, and can I use it commercially?

The source code is available, but the license is the Sustainable Use License, which is not an OSI-approved open source license. You can use the software for your own internal business purposes or non-commercial use; LICENSE.md doesn't permit providing it to others for a fee. This is not legal advice.

n8n or Make?

The official sources don't say which is better. n8n counts one full run of a workflow as one execution and can be installed on your own server. Make uses different units; see the Make guide.

Can I turn off n8n telemetry, and where is my data?

On your own server you turn it off with environment variables like N8N_DIAGNOSTICS_ENABLED=false. On Cloud, telemetry can't be turned off. Cloud data is stored in the EU, in Frankfurt; on your own server, data stays where you choose.

Tasks and setup kits that use n8n

Automating appointment reminder messages Setting up a form → spreadsheet → email automation

Related guides

Sources

Not clear in the official sources: Whether there's a Turkish interface · monthly (non-yearly) price for Cloud Starter and Pro · cancellation and refund terms for Starter and Pro · phone, age or ID requirements at sign-up · card payments from Turkey and the tax rate · minimum hardware for bare n8n · which plans include built-in MCP access · whether agencies and consultants using n8n on behalf of clients comply with the license · Startup and Education discount terms · Node.js version (the docs say 20.19-24.x, the 2.43.5 package on npm requires 24 or later) · whether Pro is 10K or 50K executions (pricing page and data docs differ) · whether n8n Assistant supports chatting in Turkish · what happens on Cloud when the monthly execution quota runs out. Each is covered in the relevant section along with what you can do.