Workflow Automation / AI Agents

n8n

The open-source workflow automation platform for AI agents

n8n is an open-source workflow automation tool that connects hundreds of apps and APIs through a visual node-based editor. In 2025 it has become the leading platform for building AI agents that take real-world actions — reading emails, updating CRMs, querying databases, and orchestrating multi-step AI pipelines.

Learn n8n with a structured course →

What n8n does

n8n lets you visually connect apps, APIs, and AI models into automated workflows called "workflows." Each step is a node — trigger, action, or transformer. It runs self-hosted (free) or on n8n Cloud, and its Code node lets developers add arbitrary JavaScript or Python logic inline. Its AI nodes let you call LLMs, build agents with tool use, and maintain memory across workflow runs.

Is it right for you?

n8n is best for…

Technical product managers who want to prototype integrations without full engineering effort
Developers building AI agent pipelines that interact with real systems
Startups that need automation power without paying Zapier enterprise prices
Teams who want full data ownership by self-hosting their automation infrastructure
Engineers who need the flexibility of custom code inside a visual workflow tool

Not the best choice if…

Non-technical users who want a no-code experience (Zapier or Make.com are simpler)
Teams with zero developer resources — n8n benefits from someone comfortable with JSON and APIs
Simple two-step automations (overkill vs. Zapier for basic triggers)

Pricing

Free

Self-hosted community edition: free forever, unlimited workflows

Paid

n8n Cloud Starter: $20/month for 2,500 executions/month

Enterprise

n8n Cloud Pro/Enterprise: from $50/month, custom execution limits, SSO, support SLA

Key features

Visual node editor

Drag-and-drop nodes for 400+ apps and services. Connect triggers to actions to conditions to AI calls in a visual canvas.

AI agent nodes

Built-in nodes for LLM calls (OpenAI, Anthropic, etc.), memory (vector stores, chat history), and tool use — build stateful AI agents visually.

Code node

Write arbitrary JavaScript or Python inside a workflow step — filter data, transform JSON, call external libraries, or implement custom business logic.

Self-hosted option

Run n8n on your own server with full data ownership. No per-execution pricing in self-hosted mode — run unlimited automations at your infrastructure cost.

Sub-workflows

Break complex automations into reusable sub-workflows that other workflows can call — essential for building maintainable agent architectures.

Error handling and retry

Built-in error handling nodes, retry on failure, and execution logs with input/output for every step — makes debugging production automations feasible.

Real-world use cases

AI email triage and response drafting

Trigger on new email → extract intent with GPT-4o → route to the right team → draft a reply with Claude → send for human approval via Slack.

Lead enrichment and CRM update

New form submission → enrich with Clearbit/Apollo → score with AI → create HubSpot deal → assign to rep → send personalized outreach email.

AI-powered content pipeline

Schedule trigger → fetch trending topics → generate blog draft with Claude → post to Notion for review → on approval, publish to CMS and share on LinkedIn.

Customer support agent

Incoming support ticket → retrieve customer history from database → answer with AI using product knowledge base → if confidence low, escalate to human.

Internal data dashboards

Scheduled trigger → query multiple data sources → aggregate with code node → generate natural language summary with AI → post to Slack digest.

How to get started with n8n

Choose: cloud or self-hosted

Start with n8n Cloud (n8n.io/cloud) for zero setup. Choose self-hosted if you need data sovereignty or unlimited executions at lower cost.

Create your first workflow

Add a Manual Trigger node. Add an HTTP Request node pointing to any public API. Connect them and click Execute Workflow.

Connect your apps

Search the node panel for apps you use (Gmail, Slack, Notion, HubSpot). Add credentials in Settings → Credentials and authorize via OAuth or API key.

Add an AI node

Add the OpenAI or Anthropic node. Connect your API key credential. Pass data from previous nodes as context to the LLM prompt.

Set a production trigger

Replace the Manual Trigger with a Webhook, Schedule, or app trigger (e.g., "Gmail: On New Email") to run the workflow automatically.

Pro tips

→Use the Code node to transform data between steps — it is much faster than adding extra app nodes just to reshape JSON.
→Pin data in the node panel to freeze test inputs so you can iterate on downstream nodes without retriggering the full workflow.
→Build sub-workflows for any logic you repeat — this is the difference between a workflow that is maintainable at 50 nodes vs. one that becomes a spaghetti nightmare.
→Use n8n's AI agent node with a memory buffer for conversational agents — it handles the chat history loop automatically.
→For production, always add a catch-all error handler node that notifies you in Slack when any execution fails.

Frequently asked questions about n8n

What is n8n and how does it compare to Zapier?

n8n is an open-source workflow automation tool that can be self-hosted and supports custom code inside workflows. Zapier is a no-code cloud SaaS with a simpler interface but higher pricing and no self-hosting option. n8n is more powerful and cheaper at scale; Zapier is faster to start with for simple two-step automations. For AI agent use cases in 2025, n8n is significantly more capable.

Do I need to know how to code to use n8n?

You can build many workflows without code using n8n's visual editor and 400+ pre-built nodes. However, to unlock its full power — custom data transformations, error handling, and complex agent logic — basic JavaScript knowledge is very helpful. Our n8n course for product teams covers exactly this: the minimum code needed to build powerful automation.

Can n8n build AI agents?

Yes, n8n has first-class support for AI agents. Its AI Agent node handles tool selection, memory, and multi-step reasoning loops using LLMs from OpenAI, Anthropic, and others. You can connect real tools (send email, query database, call APIs) as agent capabilities. Building stateful, production-ready AI agents in n8n is one of the most in-demand skills in 2025.

What is the difference between n8n Cloud and self-hosted n8n?

n8n Cloud is a managed service where Anthropic handles infrastructure, updates, and scaling. Self-hosted runs on your own server (Docker or Node.js) and is free with no execution limits — you pay only for your server. Self-hosted requires more setup but gives you full data ownership and unlimited scale at low cost. For most teams, n8n Cloud is the right starting point.

How does n8n handle secrets and credentials?

n8n stores credentials encrypted at rest in its database. In the visual editor, credentials are referenced by name rather than embedded in workflow definitions — so exporting a workflow does not expose API keys. In self-hosted deployments, you control the encryption key. n8n is also SOC 2 Type II certified for its cloud offering.

Ready to master n8n?

Build AI Agents with n8n

Build powerful AI automation workflows and agents using n8n.

View course →