What Is n8n? Pricing, Self-Hosting and LINE OA for Thai Businesses

n8n is an open-source workflow automation platform that connects LINE OA, Google Sheets, databases and APIs on a drag-and-drop canvas. The real difference from Zapier and Make is not the interface, it is the billing unit. n8n counts one run of your entire workflow as a single execution no matter how many nodes it contains, and you can self-host it for free. This guide skips the generic "what is automation" explainer and answers the three questions Thai businesses actually ask before starting: what does it cost, how well does it connect to LINE OA, and should you self-host?

What n8n Actually Is, Without the Marketing Language

There are only three concepts. Trigger: what starts things: a webhook from LINE, a schedule, a new row in Google Sheets. Node: each step after that: call an API, transform data, write to a database, ask an LLM. Execution: one full run of the workflow from start to end. Remember this word, because it is the unit n8n charges for, and everything below revolves around it.

• •Open source (fair-code): the code is on GitHub and you can install it on your own server without asking anyone
• •Genuinely self-hostable: the Community edition is free, and n8n's docs describe it as "almost the complete feature set"
• •Code where you need it: a Code node for JavaScript and Python when the visual editor runs out of road
• •Built-in AI nodes: OpenAI, Anthropic and Ollama, plus agents that call their own tools (what an MCP server is covers the standard underneath)

Where n8n Is Genuinely Cheaper: Per Run, Not Per Step

The n8n pricing page states plainly that an execution is "a single run of your entire workflow", and every plan includes unlimited steps per execution. Make and Zapier count operations or tasks, which are counted per step. Run the numbers on a job Thai shops actually have: an order arrives from LINE, the payment slip is checked, the order is written to Google Sheets, the team is notified in Slack, the customer gets a reply. That is roughly 18 nodes. At 500 orders a month, n8n charges 500 executions while a per-operation tool lands around 9,000 operations. That is not a 10-20% discount, it is a different scale, and it means the more complex the workflow, the more n8n wins. The honest other side: if your workflow is only 2-3 nodes but runs thousands of times a day, that advantage mostly evaporates. In that case re-run the comparison rather than assuming n8n is always cheaper (full three-way comparison).

What n8n Costs in 2026

The figures below are read straight off n8n's pricing page. They are the annual-billing rates, and every plan includes unlimited active workflows.

The Biggest Bill Trap Is Not Workflow Count: It Is Polling

Most people assume they get charged when a customer shows up. In reality a workflow on a Schedule trigger runs and burns an execution even when nothing happened. Do the multiplication:

• •Every 5 minutes = 8,640 executions/month, past the Starter allowance (2,500) inside the first week, and past Pro (10,000) within the month
• •Every 30 minutes = 1,440 executions/month
• •Every hour = 720 executions/month
• •Webhook = only fires on real events. A shop with 300 orders a month spends 300 executions

The good news for Thai businesses is that the LINE Messaging API is webhook-based already, so there is nothing to poll. The part to watch is Shopee and Lazada, where many setups still pull on a schedule: 15 to 30 minutes is plenty for a normal shop. Polling every 5 minutes for a "real-time" feel means paying several times more for latency no customer notices. And even when self-hosting, where executions cost nothing, aggressive polling still burns CPU and RAM on your server. It is not free either.

How Well It Connects to the Thai Stack: LINE OA, PromptPay Slips, Shopee

This is the part generic "what is n8n" articles never cover, because their authors do not work with the Thai stack.

• •LINE OA: n8n accepts LINE Messaging API webhooks directly, but LINE's reply token has a limited lifetime. Slow work like OCR or an LLM call should send a short acknowledgement first and deliver the result as a push message, rather than holding the workflow open (10 n8n + LINE OA recipes)
• •PromptPay slips: plenty of Thai customers still transfer money and send a photo of the slip into LINE. n8n takes that image, sends it to a slip-verification API, updates the order status and replies, and the whole chain counts as one execution, which is exactly why n8n's pricing model suits this kind of work
• •Shopee / Lazada: bluntly, there is no ready-made node. You use the HTTP Request node and handle request signing and token refresh yourself. That stops being no-code. If nobody on the team can read API docs, cut it from phase one
• •LINE message costs: n8n does not charge for sending messages, but LINE OA charges by plan and recipient count. An automation that broadcasts to everyone burns money on the LINE side, not the n8n side (segment before you broadcast)
• •Google Sheets: still the real database for most Thai SMEs, and n8n connects to it out of the box (connect LINE OA to Google Sheets)

Self-Hosting Is Free: But Free Up to Where?

The Community edition is genuinely free and covers almost everything, but n8n's docs list these as paid-tier only: Projects, Workflow and credential sharing, SSO (SAML, LDAP), Version control using Git, Environments, External secrets, Custom variables, Log streaming, Multi-main mode and External storage for binary data. In business terms: if you are an agency running several clients, the free edition has no Projects and no credential sharing, so everything sits in one shared workspace. No Git version control means nobody can tell who changed which workflow, or when. More importantly, n8n's own documentation marks maintenance, updates, backups and security for self-hosted deployments as "Your responsibility", versus "Handled by n8n" on Cloud.

So What Should a Thai Business Choose?

Our position is clear: most Thai SMEs should not start by self-hosting. Not because self-hosting is bad, but because the opportunity cost of your time far exceeds a Starter plan. Your hours belong to designing workflows that pay for themselves, not to chasing Docker and SSL renewals. Self-host only when one of these is true: (1) a PDPA or data-residency requirement genuinely forbids data leaving a server you control, (2) execution volume has reached the point where Cloud stops making sense, especially around the Pro-to-Business jump, or (3) someone on the team already runs servers as their actual job, not as an experiment. And before picking any tool, work out how many months your first automation takes to pay back (how to calculate automation ROI). The cheapest tool is the one you never had to buy, because the task was not worth automating yet.