n8n Community Edition: is n8n free? What is included, its limits and the real cost

"Is n8n free?" calls for a two-part answer that surprises Zapier and Make users: n8n Cloud has no permanent free plan, but the self-hosted Community Edition is free, with no limit on executions or workflows. In other words, free n8n genuinely exists — as long as you host it yourself. This guide details exactly what the free version contains, what's reserved for paid plans, what the license permits, and the costs that remain once the subscription disappears.

n8n Cloud: a free trial, not a free tier

On n8n.io, the hosted offering starts with a time-limited free trial, no credit card required — enough to test seriously, not enough to run a production automation indefinitely. After the trial, you pick a Starter, Pro, Business or Enterprise plan; we break down the per-execution billing logic in our n8n Cloud pricing guide.

That's a different positioning from Make or Zapier, which offer a permanent but narrow free tier (a few hundred tasks or operations per month). n8n's philosophy is the reverse: unlimited free usage exists, but through self-hosting — consistent with the tool's open source model, compared in detail in our n8n vs Make vs Zapier comparison.

The Community Edition: free and quota-less

The Community Edition is the version you install yourself — typically in five minutes with Docker. What it includes, with no quota or time limit:

• Unlimited workflows and executions: the only bound is your server's capacity (CPU, RAM);
• All standard nodes: hundreds of native integrations, the AI and LangChain nodes, the Code node, webhooks — nothing in the automation core is functionally crippled;
• Community nodes: freely installable, and self-hosted-only anyway;
• Encrypted credentials, basic user management, and queue mode with Redis for horizontal scaling — yes, the multi-worker production architecture is available in the free version.

Community Edition features: the exact list

The official page on Community Edition features, checked on October 14, 2026, lists what the free edition does not include: custom variables, environments, external secrets, external storage for binary data, log streaming, multi-main mode, projects, SSO (SAML, LDAP), sharing of workflows and credentials, and version control using Git. Logging and queue mode are included.

There is also a middle step few people know about: the registered Community Edition. Enter an email address in the instance settings and n8n sends a free license key that unlocks three features: folders to organize workflows, debug in editor (copy and pin execution data while you work on a workflow) and custom execution data (save, search and annotate execution metadata). No payment, no execution quota.

What this means in practice: the free edition is a complete engine for one person or one team sharing a single admin account; what it lacks is everything that separates people (projects, sharing, SSO) and environments. If you need those, the next step is the Business license (€800 per month, €667 billed annually, self-hosted only, prices as of October 14, 2026, detailed in our n8n pricing guide). If you don't, the real cost is the server and the upkeep: the FlowKit Self-Hosting Kit (€79, one-time) is a tested Community Edition production stack (PostgreSQL 18, HTTPS, queue mode, verified encrypted backups, update with rollback).

For a technical team automating its own processes, the Community Edition covers the vast majority of needs. That's the scenario we cost out in our analysis of the real cost of self-hosted n8n.

Rate limits and other limits on self-hosted n8n

The self-hosted Community Edition has no execution quota and no built-in rate limiter. What does exist, from the official documentation (concurrency control, endpoint variables), checked on October 13, 2026:

Limit	Default on self-hosted	How to change it
Executions per month	none	—
Concurrent production executions (regular mode)	no limit: "Concurrency control is disabled by default"	N8N_CONCURRENCY_PRODUCTION_LIMIT=20 (any value other than -1 turns it on; extra executions wait in a queue)
Concurrent jobs per worker (queue mode)	the worker's --concurrency flag	--concurrency, or N8N_CONCURRENCY_PRODUCTION_LIMIT, which takes precedence when set
Maximum payload size	16 MiB (N8N_PAYLOAD_SIZE_MAX)	raise it, at the cost of more memory and CPU
Files in form-data webhook payloads	200 MiB (N8N_FORMDATA_FILE_SIZE_MAX)	same variable

Turning the concurrency limit on is the main protection a small server has: without it, a burst of webhooks can launch dozens of executions at once and exhaust RAM. For heavier loads, queue mode with Redis spreads executions across workers.

The rate limits you will actually hit are those of the services your workflows call: OpenAI, Google, Slack or a CRM answer 429 Too Many Requests well before n8n complains. Retries, batching and retry-after handling are covered in our guide to AI API rate limits in n8n.

429 errors: what the official n8n documentation recommends

The page Handle rate limits of the n8n documentation (checked on October 17, 2026) is the reference for the self-hosted Community Edition as much as for Cloud, since the limit comes from the API, not from n8n. When a node receives a 429 Too Many Requests, it errors and the output panel shows "The service is receiving too many requests from you" next to the service's own message; the quota itself is listed in that service's API documentation. Three settings deal with it, all available in the free edition:

Where	Setting	What it does
Any node, Settings tab	Retry On Fail, with Wait Between Tries (ms)	retries after a pause; set the wait above the API's limit (docs example: one request per second → 1000)
Before / after the node that calls the API	Loop Over Items + Wait node, looped back	splits the items into small batches and pauses between them
HTTP Request node, Add Option	Batching: Items per Batch, Batch Interval (ms); Pagination	the built-in equivalent of Loop Over Items + Wait; pagination for large responses

What this looks like in the workflows we sell: among the 12 workflows of the FlowKit packs, two already pace their calls with Loop Over Items, one item at a time — the Notion → vector store sync (one page, hence one round of OpenAI embeddings, per pass) and the automatic reminders (one GPT-4o-mini email per incomplete file). Neither has Retry On Fail turned on nor a Wait node: on a large Notion base or a long list of files, they are the first two settings to add if OpenAI starts answering 429. Retry patterns by status code (retry a 429, never a 401) are detailed in our guide to HTTP Request retry and timeout, and alerting on failed executions in our Error Workflow guide.

On the instance side, the FlowKit Self-Hosting Kit ships queue mode with a per-worker concurrency setting (WORKER_CONCURRENCY, 10 in its .env.example) and verified encrypted backups; if you would rather have someone else set up and tune the instance, you can also hand it to a professional.

For comparison, n8n Cloud does cap usage per plan. The pricing page listed, on October 13, 2026 (annual billing): Starter €20/month for 2,500 executions and 5 concurrent executions; Pro €50/month for 10,000 executions and 20 to 50 concurrent executions; Business €667/month for 40,000 executions, self-hosted; Enterprise on quote. On Cloud, executions beyond the concurrency limit queue and run in order when capacity frees up.

What's reserved for paid plans

Community Edition restrictions aren't about volume but about governance — features that matter mostly when several teams share the instance:

• SSO (SAML/OIDC) and LDAP to plug in enterprise authentication;
• Projects, roles and granular permissions: in the Community Edition, fine-grained control over who sees and edits what remains basic — our guide to users and permissions on self-hosted n8n details what's doable without paying;
• Separate environments with built-in Git deployment: the native dev/staging/prod feature is paid, but an equivalent pipeline can be built by hand — see our dev/prod environments and versioning workflows with Git guides;
• Centralized variables, external secrets (Vault, AWS Secrets Manager), log streaming to a SIEM;
• Vendor support with an SLA.

The exact list evolves across releases: n8n.io/pricing and the official documentation are the source of truth. The stable part is the logic: the engine is free, the enterprise tooling is paid.

The Sustainable Use License: what "free" allows

n8n isn't under a classic open source license (MIT, Apache) but under the Sustainable Use License, a fair-code license. Its license file sets the rule in one sentence: "You may use or modify the software only for your own internal business purposes or for non-commercial or personal use", and distribution is allowed "only if you do so free of charge for non-commercial purposes". Files with .ee in their name or folder are excluded and need an n8n Enterprise License. In practice:

• Allowed for free: any internal use, including commercial — automating your company's processes, or your clients' if you're a consultant or agency (with each client hosting or owning their instance);
• Forbidden without a commercial agreement: offering n8n itself as a hosted service to third parties, reselling it, or white-labeling it inside your SaaS product (n8n has a dedicated embed offering for that).

This hybrid model isn't an n8n quirk: the reference economic analysis by Josh Lerner and Jean Tirole, "Some Simple Economics of Open Source", published in 2002 in the Journal of Industrial Economics (see the study), already showed that open-code projects stay viable through value-capture mechanisms complementary to the free code — services, hosting, enterprise features. The Sustainable Use License is exactly that: it guarantees your free internal use is sustainable because the vendor funds itself elsewhere.

The real costs of a "free" n8n

Removing the subscription doesn't remove the bill, it changes its shape:

• The server: a VPS is enough to start, sized to your workflows (AI workflows and file processing are RAM-hungry);
• Human time: installation, updates that don't break things, PostgreSQL backups, monitoring, incident recovery. This is the most underestimated line item, and often the most expensive;
• Third-party APIs: free on the n8n side doesn't mean free on the OpenAI or Anthropic side — AI calls are still billed per token by their provider, worth tracking workflow by workflow.

The free-vs-paid choice is therefore less "save money or not" than "pay in subscription or in admin time". Our self-hosted vs Cloud comparison gives the full decision grid by profile.

In short

Free n8n means the self-hosted Community Edition: unlimited executions and workflows, all nodes, queue mode — a complete automation engine whose only bounds are your server and your admin time. n8n Cloud can be tried for free but is paid per use. If you're starting out: take the Cloud trial to validate your use cases, then make an honest trade-off between subscription and self-hosting once you know your volume — and if the answer is self-hosting, our Docker installation guide is the place to start. For a first real use case on a free Community Edition, AI email sorting over IMAP only relies on standard nodes, available without a paid licence.