Descriptions:
Nate Herk walks through the complete delivery lifecycle for client-facing AI workflow projects built on n8n, covering the questions he says come up most often: who should host the instance, how to handle API keys securely, how to test before go-live, and what a professional handover looks like. The video is organized around n8n’s licensing boundaries, which Herk explains in plain terms—the free and sustainable-use license allows building workflows for a single business internally, but turning n8n into a hosted platform for multiple clients requires a commercial or enterprise license.
Three hosting models are compared: client-owned n8n instances (recommended for most agency work), agency-hosted infrastructure used only for internal operations or white-label service delivery, and full SaaS products built on n8n (which require commercial licensing). On API key management, Herk recommends clients own and pay for their own keys, with secure transfer handled via tools like 1Password or encrypted one-time-link sharing rather than passing credentials over Slack or email.
The second half covers testing and handover: collecting real anonymized sample data before signing a contract, defining success and failure criteria upfront, and deliberately stress-testing workflows with bad, missing, or duplicate data before production. The video closes with a walkthrough of one of Herk’s earliest sold automations as a concrete reference point. It is an unusually complete operations guide for freelancers and agencies building commercial AI automation services.
📺 Source: Nate Herk · Published December 27, 2025
🏷️ Format: Workflow Case Study
