Rating certificate · AW/26-09/85526 Sept 2026
68of 100
The Register / Agent-building platforms / n8n
Register entry · Agent-building platforms · updated 26 Sept 2026
n8n
n8n is a node-based workflow automation tool from Berlin-based n8n GmbH. Workflows combine app integrations, JavaScript/Python code steps and LLM agent nodes, and can run on n8n Cloud or be self-hosted from the GitHub repository. The code is published under the Sustainable Use License (fair-code, not OSI-approved) with separate enterprise-licensed features. The company raised a $180M Series C at a $2.5B valuation in October 2025.
What reviewers and users praise
- Practitioner reviewers describe it as good value once past the learning curve, especially self-hosted (Softomate, Osher Digital)
- Community users report it working well for small AI automations at low cost (Hacker News)
- Seen by developers as a flexible middle ground between no-code and custom code (Hacker News, Reddit)
What they criticise
- Not beginner-friendly; several reviewers and users cite a multi-hour learning curve and difficulty for non-coders
- Repeated critical security advisories (RCE and sandbox escapes) in the last 24 months
- Some developers now prefer generating code with coding agents over building visual flows (Hacker News, Reddit)
Experts vs the crowdcrowd cooler by 14
Experts67
Crowd53
| Price | Free |
|---|---|
| Use it via | Web app, API, MCP server, Desktop app |
| Runs on | web, linux, macos, windows |
| Licence | LicenseRef-n8n-Sustainable-Use-License |
| Updated | Evidence collected 26 Sept 2026 |
How the rating was worked out
Weights for Agent-building platforms: adoption 35%, experts 35%, crowd 30%. Pillars without enough evidence are left out and the others re-weighted. The method.
| Adoption & momentumweight 35% | 83 | GitHub stars 206k → 100 · npm weekly downloads 88k → 53 · 135 releases in 90 days → 100reach × 0.7 + momentum × 0.3 |
|---|---|---|
| Expertsindependent reviews · weight 35% | 67 | 3 independent reviews (2 positive · 1 mixed · 0 negative), averaged toward 50 for small samples. |
| Crowdcommunity sample · weight 30% | 53 | 11 dated posts from hackernews, reddit: 5 positive · 2 mixed · 4 negative (positive + ½ mixed) ÷ n, averaged toward 50 for small samples |
| Rating | 68 | medium confidence (3 of 3 pillars, 11 community posts). |
Editorial adjustment. Incident-history re-grade under methodology revision 2 (see the trust section).
Trust & safety
Seven dimensions graded A–D from documented facts. Overall: B.
C
Permission model
Workflows execute automatically on their triggers; AI Agent tool calls can be gated with human-in-the-loop approval steps (enhanced in 2.35 for Slack/Telegram), but approval is opt-in per workflow. Code nodes run in task runners that admins can harden. source
A
Data access scope
Access is scoped by credentials per integration, RBAC projects, and admins can block specific nodes; OAuth scope display for managed credentials added in 2.35.4. source
A
Data storage
Fully self-hostable (including air-gapped); n8n Cloud optional. source
A
Data retention & training
Privacy docs: context sent to LLMs for AI features 'will not be incorporated, used, or retained to train the models of the AI services. Any data will be deleted after 30 days.' Cloud session recordings deleted after 21 days; self-hosters control execution pruning. source
C
Incident history (24 months)
196 GitHub security advisories published 2024-09-26 to 2026-09-26, 22 rated critical, many RCE or sandbox escapes. All listed advisories have patched versions, but critical issues recurred through 2026. Regraded from D: Revision 2: serious issues that were each fixed and publicly disclosed grade C; D is kept for unresolved issues or slow or undisclosed fixes. source
A
Compliance
Trust center lists SOC 2 Type 2 and SOC 3 reports and GDPR. source
B
Transparency
Source code public on GitHub under the Sustainable Use License (fair-code, not OSI-approved) plus enterprise-licensed files; detailed public release notes. source
The experts
4 named independent reviews found; dated reviews from the last 12 months are scored.
- positive
n8n is the best-value automation platform in 2026. n8n is not beginner-friendly - expect a 4-10 hour learning curve
Softomate Solutions · 6 Jun 2026 · softomatesolutions.com - positive
After 8 months and 12 client projects, n8n has become my go-to tool for custom AI automation. It's not perfect, but it hits a sweet spot
Nova (dev.to) · 1 Apr 2026 · dev.to - mixed
The 2026 verdict: Zapier for small teams without a technical owner; n8n once you cross a few thousand executions per month or need self-hosting.
Osher Digital · 1 Jul 2026 · osher.com.au - mixed
Verdict: n8n is worth it if you're willing to invest 5-10 hours getting comfortable with its logic.
FlowBuilderHQ · — · flowbuilderhq.com
The crowd
11 coded posts from hackernews, reddit.
5 positive2 mixed4 negative
- positive
I have hacked together a basic email classification with n8n and cheap LLM model, and it works awesomely and costs peanuts.
An update on leaving Gmail for Fastmail · 17 Aug 2026 · news.ycombinator.com - negative
After that, I tried very hard with n8n but just found it impossible to work with
Show HN: Ballet – Workflow automation that writes integrations against any API · 13 Aug 2026 · news.ycombinator.com - negative
But the llm can also program you the exact automation you want! Much more efficiently than gui madness with N8N.
Claude Code Routines · 14 Apr 2026 · news.ycombinator.com - negative
oh wow ive been looking for n8n cloud observability for so long. its so annoying that n8n hasnt done anything regarding this.
Bringing Observability to Your N8n Workflows: OpenTelemetry in Action · 12 Apr 2026 · news.ycombinator.com - positive
I've used n8n for years and told everyone who would listen to give it a try, well before LLMs.
OpenClaw surpasses React to become the most-starred software project on GitHub · 2 Mar 2026 · news.ycombinator.com - positive
Ultimately I think the agent implementation in n8n is better done
OpenClaw – Moltbot Renamed Again · 30 Jan 2026 · news.ycombinator.com - positive
n8n allows you to work at a higher level, and working at this level allows you to do things in a way that's more likely to be "correct".
RIP Low-Code 2014-2025 · 27 Jan 2026 · news.ycombinator.com - negative
This is, to me, what sinks most of those services like Zapier and n8n in power-user market
Show HN: Ballet – Workflow automation that writes integrations against any API · 13 Aug 2026 · news.ycombinator.com
Pricing
As published, checked 26 Sept 2026.
| Plan | Price | Notes |
|---|---|---|
| Community Edition (self-hosted) | Free | Self-hosted from GitHub under the Sustainable Use License. |
| Starter (cloud) | Custom | EUR 20/mo billed annually; 2.5K workflow executions, 5 concurrent executions, unlimited users. |
| Pro (cloud) | Custom | EUR 50/mo billed annually; 10K workflow executions, 20 concurrent executions. |
| Business (self-hosted) | Custom | EUR 667/mo billed annually; 40K executions, SSO, Git version control. |
| Enterprise | Custom | Contact sales. |
Recent changes
- 15 Sept 2026
n8n 2.40: Microsoft Dataverse core node plus 31 other features. source
- 25 Aug 2026
n8n 2.37: streaming responses for Message an Agent. source
- 11 Aug 2026
n8n 2.35: enhanced human-in-the-loop approval for Slack and Telegram. source
- 21 Jul 2026
n8n 2.32: scope selection for MCP OAuth consent. source
- 28 Jul 2026
n8n 2.33: redesigned instance settings for n8n Assistant. source
Sources
Every figure above traces to one of these, observed up to 26 Sept 2026.