How can I host n8n for free?

Sign up at flowengine.cloud and deploy a free n8n instance in under 30 seconds. No credit card required for the free tier.

How do I build an n8n workflow with AI?

Visit flowengine.cloud/chat, describe your automation in plain English, and FlowEngine generates the n8n workflow JSON for you to import or deploy.

What makes FlowEngine different from n8n Cloud?

FlowEngine adds an AI workflow builder, a free workflow analyzer, and 100+ bundled LLM models without per-model API keys, on top of standard managed n8n hosting. Pricing starts at $7 per month versus $20 for n8n Cloud, and a free tier is available.

What LLM models are available inside FlowEngine?

Over 100 models including GPT-4o, Claude 3.5 Sonnet, Claude 3 Opus, Gemini 1.5 Pro, Llama 3 70B, and Mistral Large. Accessible inside any n8n workflow without managing separate API keys.

Does FlowEngine support custom domains?

Yes. All paid plans include custom domain support with automatic SSL provisioning via Let's Encrypt.

Is the n8n workflow analyzer really free?

Yes. flowengine.cloud/n8n-workflow-analyzer is completely free, requires no signup, and runs in-browser so workflow contents stay private.

Can I export my workflows out of FlowEngine?

Yes. Workflows are stored in standard n8n JSON format and can be exported from the n8n UI at any time. There is no vendor lock-in beyond the hosting layer.

Introducing FlowEngine Functions: Code That Bills by the Second

You hit the wall the moment a workflow needs real code. The n8n Code node is great for quick snippets but caps fast: no native libraries, no long-running jobs, no real Python. The serverless alternatives all live outside your workflow surface, on a different cloud, with a different CLI, a different invoice, a different dashboard. Meanwhile your n8n graph sits there, half done, waiting for the one custom function it actually needs.

Today FlowEngine ships Functions: serverless code execution that lives inside your n8n and agent workflows. Any language Buildpacks supports. Deploy in seconds. Scales to zero when nobody's calling. Charges only when it runs.

What It Does

A Function is a folder. You drop in your code, click Deploy, and FlowEngine hands back a URL. That URL is HTTP-callable from anywhere, including the HTTP Request node inside your n8n workflows, your OpenClaw agents, your Slack bots, or your phone.

Any language: Python, Node, Go, Java, Ruby, PHP, .NET. Buildpacks autodetects the runtime. No Dockerfile required.
Triggers built in: webhook URL by default, plus Slack mention or DM, plus cron schedules. Pick when you create the function.
OAuth handled for you: Composio pass-through means Gmail, Google Calendar, Notion, GitHub, and 250 other providers are one click. The token shows up as an env var. You write the business logic, not the OAuth dance.
Scales to zero: idle functions cost nothing. Cold starts are fast enough for n8n workflow steps.
Logs and runs visible in the portal: every invocation streams to the portal in real time so you can debug without SSH.

How to Use It

Open Functions in the portal and click New.
Pick a template (Morning Brief, WhatsApp Chatbot, Chromium browser) or start from a blank folder.
Edit the code in the browser or push from your local repo via the CLI.
Pick a trigger: webhook URL, Slack workspace mention, or cron schedule.
Click Deploy. The build runs through Buildpacks and you get back a live URL in a few seconds.
Drop that URL into any n8n HTTP Request node and your workflow now has a custom code step that runs on your stack.

Why We Built It

The same conversation kept happening with FlowEngine users. The workflow was almost done in n8n, but one node needed Python (the SDK exists in Python, not JavaScript) or one needed a real headless browser or one needed a long-running background job. Each user ended up with a side stack: a stray Modal app, a one-off Cloud Run service, a Lambda nobody wanted to maintain. The n8n graph was clean and visual. The custom code lived in a different tab, billed on a different invoice, monitored in a different dashboard.

Functions collapses that back into one surface. Your workflow lives in n8n. Your custom code lives in FlowEngine Functions. Both deploy from the same portal. Both bill on the same workspace. Logs land in the same place. When the n8n graph calls the Function, you see the request hit a Functions log line one second later. That visibility was the missing piece.

For agency teams running custom logic for every client, Functions removes the deployment-target problem entirely. Spin up a per-client function, bind it to that client's n8n instance, ship it the same day. No new infra, no separate cloud account, no extra CLI to teach the team.

And if you're building production AI workflows with OpenClaw, Functions gives you the deterministic compute layer agents call out to: the same input produces the same output, the function version is pinned, the run is auditable. That's the difference between "demo AI" and AI you can actually ship to a regulated customer.

What's Next

Functions launches with templates for Morning Brief (calendar plus inbox plus Slack summary), WhatsApp chatbot, headless Chromium scraper, and a blank starter for each language. The CLI ships alongside for local development and CI deploys. Next iteration locks in Slack as a first-class trigger type plus deeper Composio provider coverage.

Try It

Functions is live now in the FlowEngine portal for every workspace. Open /portal/functions to deploy your first one. Billing details and tier specifics live on the pricing page. If you want a walkthrough, book a demo and we will show you a real workflow from problem to deploy in under ten minutes.