# n8n or a platform? How agencies should deliver AI agents

> n8n or a platform for an agency delivering WhatsApp AI agents: when n8n is the best choice, what its license allows, the cost as your client base grows and how to migrate flows.

- **HTML version:** https://tolky.to/en/blog/n8n-or-platform-for-agencies
- **Author:** [Marlos Carmo](https://tolky.to/en/author/marlos), Co-founder of Tolky
- **Published:** November 20, 2026
- **Reviewed:** November 20, 2026
- **Language:** en

## Summary

For an agency, **n8n** is the best choice when the AI agent is one piece of a larger automation, the client base is small and the client only receives the output; the license allows running workflows for clients, but not letting clients into the instance or reselling n8n under another brand. As the client base grows and each client wants to see conversations, answer on the same number and track the funnel, a platform with an inbox, CRM and per-client access costs less in hours. Tolky Flows import n8n JSON, and white-label starts from the Enterprise plan's Your brand and domain feature.

The first AI agent in n8n ships over a weekend. It works, the client is happy, the agency closes two more. By the tenth client, someone on the team spends the morning re-subscribing webhooks, swapping expired tokens and explaining to a client why the agent replied twice.

If you run an agency, you did not build it to become server support. The right question is not whether n8n is good (it is), but which delivery model your client base can carry a year from now. This guide compares n8n and a platform from the agency's point of view: license, cost in hours, client access and migration.

*Updated in November 2026.*

## What is the difference between delivering with n8n and with a platform?

**Delivering with n8n means building and hosting, for each client, a workflow that connects the WhatsApp API to an AI model; delivering with a platform means configuring an agent inside a product that already includes an inbox, CRM, human handoff and reports.** In the first case, the agency owns the infrastructure. In the second, it owns the configuration and the operations.

The technical walkthrough for the first path is in the guide to a [WhatsApp AI agent with n8n](https://tolky.to/en/blog/whatsapp-ai-agent-n8n). Here the topic is what changes when the same build repeats across ten or twenty clients.

## When is n8n the best choice for an agency?

**n8n is the best choice when WhatsApp is one piece of a larger automation, the agency has a dedicated technical person and the client wants to receive an outcome, not run a tool.** In those cases, n8n's flexibility is worth more than everything a platform includes out of the box.

Situations where n8n wins:

- **Back-office automation.** Spreadsheet, ERP, email and WhatsApp connected in a workflow nobody on the client side needs to open.
- **Few clients, each with a very different workflow.** Customization over standardization.
- **Validating a use case.** Testing in days whether an agent solves the problem before selling a bigger project.

Outside those cases, the cost shows up in customer service: the client wants to see conversations and answer on the same number.

## What does the n8n license allow an agency to do?

**The n8n license lets an agency build, run and maintain workflows for clients on its own instance and charge for it, as long as clients do not access the instance.** What it does not allow is hosting n8n as a service for clients to build workflows or reselling n8n under another brand.

n8n uses the Sustainable Use License, which limits use to "your own internal business purposes" ([license text](https://github.com/n8n-io/n8n/blob/master/LICENSE.md)). The [license FAQ](https://docs.n8n.io/n8n-community-license/community-license/license-faq), checked on November 20, 2026, covers the agency case:

- **Allowed:** delivering automations to multiple clients as a managed service, provided they do not access your instance directly, and charging for workflow creation, setup, maintenance and consulting.
- **Not allowed:** hosting n8n as a service and letting clients build workflows, or letting end users build or configure workflows through your product.
- **Not allowed:** white-labeling n8n by removing its branding and offering n8n as is.

In practice, an agency that delivers with n8n sells outcomes, never access. If the plan is to let clients build workflows in an agency-branded screen, the FAQ points to n8n's sales team for an Enterprise license. When in doubt, ask n8n and your lawyer; this post is not legal advice.

## What gets heavier as the client base grows?

**As the client base grows, whatever multiplies per client gets heavier: maintenance, WhatsApp numbers, history and access.** One workflow is work once; twenty workflows are work every day.

### Maintenance that multiplies

Each client has its own workflow, credentials, token, webhook subscription and n8n version to update. A change that applies to everyone is repeated in every workflow, and an incident at 10:00 p.m. needs someone on call. When only one person understands the workflows, that person becomes the bottleneck for the whole client base.

### Official WhatsApp API per client

Each client has its own number, Meta account and templates, with quality rating, messaging limits and fees to track. Since January 15, 2026, the WhatsApp Business Platform terms bar general-purpose AI assistants; agents that serve a company's own customers remain allowed, as the post on [WhatsApp's rule for AI chatbots](https://tolky.to/en/blog/whatsapp-bans-general-purpose-ai-chatbots) explains. Shortcuts through WhatsApp Web put the number at risk; see [unofficial WhatsApp API](https://tolky.to/en/blog/unofficial-whatsapp-api-risks).

### History, CRM and client access

The client wants to read conversations, their team wants to take over when the agent hands off and the manager wants to see opportunities. n8n has no inbox or CRM, and the license does not let the client into the instance, so the agency ends up building a second tool.

### Personal data on your infrastructure

By hosting the workflows, the agency processes personal data on the client's behalf and acts as a [processor](https://tolky.to/en/glossary/data-processor) under data protection laws such as Brazil's LGPD: server, execution logs, backups and whatever goes to the AI model need access control and records.

## What does it cost: maintenance hours versus a license

**For an agency, the cost of n8n lies less in the subscription and more in maintenance hours, which grow with every client; a platform's cost is a per-client license, which the client can pay directly or within the retainer.** Compare both lines at the client base size you want to reach.

n8n Cloud has plans starting at €20 a month billed annually, with 2,500 executions, and the self-hosted Community Edition is free ([n8n pricing](https://n8n.io/pricing/), checked on November 20, 2026). The subscription is rarely the problem. What grows is technical time:

| Client base (illustrative) | Maintenance hours per month | Cost in hours, at US$ 40 an hour | Who covers on-call |
| :--- | :--- | :--- | :--- |
| 3 clients, 3 h each | 9 h | US$ 360 | The technical partner |
| 10 clients, 3 h each | 30 h | US$ 1,200 | One dedicated person |
| 25 clients, 3 h each | 75 h | US$ 3,000 | A team, on a rotation |

Hours and hourly rate are made up to show the math; measure yours over a real month. AI model, server and Meta fees are left out. On a platform, infrastructure hours disappear, but operating hours stay: reviewing conversations, updating the knowledge base and running campaigns. Those are the hours that become a retainer, as the guide to [how much to charge for a WhatsApp AI agent](https://tolky.to/en/blog/how-much-to-charge-for-whatsapp-ai-agent) shows.

## n8n, a vendor-branded platform or white-label?

**There are three ways to deliver AI agents to your client base: n8n on your own, a platform under the vendor's brand or a white-label platform under the agency's brand.** The difference is who maintains the infrastructure, what the client accesses and whose brand it is.

| Criterion | n8n on your own | Vendor-branded platform | White-label platform |
| :--- | :--- | :--- | :--- |
| Who maintains servers and updates | The agency | The vendor | The vendor |
| Inbox for the client's team | None; needs another tool | Yes | Yes |
| CRM and contact history | Built separately | Yes | Yes |
| What the client accesses | Only the output (license rule) | Their own account, under the vendor's brand | Their own account, under the agency's brand and domain |
| Who bills the client | The agency | The vendor or the agency, depending on the model | The agency |
| Best for | Few clients and back-office automation | Agencies that want to sell services without running a product | Agencies that want their own product |

The detailed white-label comparison is in the guide to the [white-label AI agent](https://tolky.to/en/blog/white-label-ai-agent-for-agencies). To see how the same dilemma shows up when building a whole operation, read [AI agency: what it is and how to start one](https://tolky.to/en/blog/ai-automation-agency-what-it-is-how-to-start).

## What does this change for your agency?

**Choosing between n8n and a platform is a business model decision: with n8n, the agency sells engineering; with a platform, it sells operations and relationships.** Both make money, but they scale differently.

- **Decide for the client base you will have a year from now,** not for the first client. What works for three clients can stall at twenty.
- **Charge for operations, not servers.** Clients pay better for conversation reviews, campaigns and reports than for hosting.
- **Standardize a base agent per industry** and adapt it per client; that is easier on a platform that keeps versions and simulations.
- **Keep n8n where it shines.** Back-office workflows can stay in n8n, triggered by webhook, with no client access to the instance.
- **Decide whose brand it is.** If the client should see the agency from login to system email, the path is white-label.

## How do you migrate from n8n without losing your workflows?

**Tolky Flows import an n8n workflow's JSON for review in the visual editor and export flows in n8n format, with caveats for steps that have no equivalent.** The agency does not redraw its client base from scratch.

The path, described on the [n8n integration](https://tolky.to/en/integrations/n8n) page:

1. **Export each n8n workflow as JSON.**
2. **In Flows, use Import and choose the file.**
3. **Review the nodes and publish.** Steps without an equivalent are rebuilt, usually with an HTTP request or advanced code.

What were loose pieces in n8n come ready: a native message trigger, an AI agent configured by mission with the account's knowledge base, handoff to a person with the history and CRM steps. Before publishing, the agency tests the agent in simulations and keeps versions ([trustworthy AI](https://tolky.to/en/trustworthy-ai)). For other systems, there is the Integration node with more than 1,000 apps, HTTP requests, webhooks, a REST API and MCP ([integrations](https://tolky.to/en/integrations)). See also [WhatsApp automation flows](https://tolky.to/en/features/whatsapp-automation-flows).

![Real screenshot of Tolky's Conversations screen in English, in the List view, with the All, Unread and Paused filters and a list of WhatsApp chats from several contacts, each with name, time and a summary of the request, in the inbox where the team takes over when the agent hands off](https://tolky.to/screenshots/conversas-whatsapp-en.png)

The screenshot shows what n8n does not include: the inbox where the client's team takes over conversations, following the rules for [AI to human handoff](https://tolky.to/en/blog/ai-to-human-handoff-7-rules). For agencies that want their own brand, the Enterprise plan's Your brand and domain feature puts the platform on the agency's domain, with logo, colors, login screen and system emails, and the agency enters every client account with a single login.

**What Tolky does not do:** the import does not convert every n8n node, there is no ready-made Zapier or Make connector (the bridge is webhooks and HTTP requests), and Instagram and SMS are coming soon. WhatsApp starts on the Starter plan, and Meta's fees are not part of the plan; see the [plans](https://tolky.to/en/pricing).

## Frequently asked questions

### Can an agency sell automation built in n8n?

Yes. According to the license FAQ, checked on November 20, 2026, an agency can build, run and maintain workflows for clients on its own instance and charge for creation, setup and maintenance. The client receives the output without accessing the instance.

### Can I resell n8n under my agency's brand?

Not under the standard license. The n8n FAQ prohibits white-labeling n8n and hosting it for clients to build workflows. For that case, n8n points to its sales team.

### GoHighLevel or n8n: which is better for an agency?

They serve different purposes. n8n is an automation tool for the agency to operate; GoHighLevel is a platform for agencies with CRM and a white-label option. Compare platforms on the [alternatives](https://tolky.to/en/alternatives) page before you decide.

### Can you use n8n and a platform together?

Yes. Many agencies keep customer service on the platform and back-office workflows in n8n, triggered by webhook or HTTP request. The client uses the inbox and CRM, and n8n stays an internal tool.

### When does it make sense to stay on n8n only?

When the client base is small, workflows differ a lot from one another and no client needs a team answering on the same number. In that scenario, n8n's flexibility makes up for the maintenance.

---

## Conclusion

n8n is the best choice for validating use cases and for back-office automations the client never needs to open. As the client base grows and each client wants conversations, a team, CRM and reports, maintenance hours start costing more than a platform, and migration reuses the workflows you already have.

## Your agency can deliver this with Tolky

**The Tolky Partner Program is for agencies that want to bring AI agents, CRM and WhatsApp campaigns to their own client base, with recurring earnings.** On the White-label track, your agency runs the platform under its own brand and domain, sets the final price and bills its clients, without maintaining servers. Training and certification are free for program members, and the Tolky team supports every initial project in 1:1 sessions.

[Explore the Partner Program →](https://tolky.to/en/partners) Let's talk?

---

This is the Markdown version of the page. The full version, with images and navigation, is at https://tolky.to/en/blog/n8n-or-platform-for-agencies

Tolky: https://tolky.to
