Video Editing, AI Content Workflows, and Privacy-First Systems for Founder-Led Teams.
Nomadic Filmworks helps founder-led businesses, remote-first companies, and small B2B teams in Europe communicate more clearly, reuse content more effectively, and work more independently through documentary-style video editing, practical AI workflows, automation, and self-hosted infrastructure.
Four building blocks. One coherent service.
Each of these areas has its own depth — but they are designed to work together. Most projects combine two or more.
Cinematic Video Editing
Documentary-driven editing that gives your story structure, rhythm, and emotional clarity — from raw footage to a finished piece people actually want to watch.
More about editingAI-assisted Content
From one core piece, we build a reusable content pipeline: clips, snippets, subtitles, and text blocks — using AI where it clearly saves time without losing quality.
More about AI contentVideo Avatars & Interfaces
Video avatars and interactive formats that make your knowledge more accessible — for onboarding, training, FAQs, or as entry points into deeper content.
More about avatarsPrivacy-first Stack Setup
Self-hosted infrastructure on EU servers — Nextcloud for files, n8n for automations, GDPR-oriented delivery. Away from scattered SaaS, toward more operational control.
More about the stackNot four separate services. One production logic.
Most founder-led teams don't need more tools or more content. They need their existing expertise made visible, their content made reusable, and their workflows made independent.
That's why I combine manual editing, AI workflows, avatar formats, and self-hosted infrastructure — not as separate businesses, but as connected parts of the same delivery.
Editing makes quality visible
Documentary-style editing turns raw material into communication assets that actually show what your business can do — not just tell.
AI workflows make content reusable
Instead of one-off production, a smart pipeline turns each core piece into clips, variants, and formats that keep working after delivery.
Avatars & interfaces make content accessible
Video avatars, interactive explainers, and chat-based entry points turn expert knowledge into formats that scale beyond the original recording.
Self-hosted infrastructure creates control
EU-hosted files, n8n automations, and GDPR-oriented delivery mean your content and workflows don't depend on scattered US-based SaaS platforms.
This is not a demo setup — it's how I run my own operations. Every tool is self-hosted in Germany, GDPR-oriented by design, and connected through n8n automations I built and maintain myself. I build the same kind of setup for clients.
Two concrete starting points.
Both are scoped for small teams and solo founders. Fixed deliverables, fixed timelines, clear prices. No open-ended retainer to start — just a defined project you can evaluate.
GDPR-Ready Operations Stack
For solo founders and small teams (2–5 people) that want to move off Google Drive, Notion, or other US-based SaaS — onto infrastructure they actually control. Your data stays in Germany. The setup is documented so you can maintain it yourself.
- Server setup on Hetzner (Nuremberg, Germany)
- Nextcloud — GDPR-oriented file storage and team collaboration
- Baserow — open-source database, replaces Airtable or Notion tables
- 1 custom n8n automation of your choice (e.g. intake form → database, email trigger, file routing)
- Handover documentation so you can maintain and extend it independently
- 30-min walkthrough call at delivery
Best fit when: you're currently using tools you don't fully control, GDPR compliance is a real concern, or you're spending too much on SaaS subscriptions that don't fit your workflow.
Ask about the Starter StackCustom Workflow Automation + Infrastructure
For small B2B teams (5–20 people) with a concrete operational problem: too much manual coordination, scattered tools, GDPR risk, or workflows that don't connect. I analyse the friction, build the infrastructure, and hand it over with documentation.
- Everything in the Starter Stack
- 2–3 custom n8n automations built around your actual processes (e.g. client intake → Baserow → email notification, content delivery pipeline, approval workflows)
- Nextcloud + Collabora for document collaboration without Google
- Baserow configured as a lightweight CRM or project tracker
- Full handover documentation + 60-min team training call
- 30-day async support window after delivery
Best fit when: your team is losing hours to manual tasks that could be automated, your current tools don't talk to each other, or you need GDPR-oriented delivery infrastructure you can trust.
Ask about the Operations SprintEvery inquiry lands in your own database automatically — no Gmail tab-switching, no copy-paste into spreadsheets.
A new client triggers the entire onboarding chain automatically — file structure ready, CRM updated, welcome sent. No manual steps.
Real work. Visible quality.
Real examples from real projects — showing both how client delivery works and how production judgment applies to newer tools like AI video.
90-second promo teaser for group travel in Colombia
KaizenTravel needed a promo video that would make their Colombia group trips feel tangible and worth joining — not another slideshow of postcard images.
From around four hours of professionally shot RAW footage, I selected the strongest scenes and shaped them into a focused 90-second story built around atmosphere, movement, and the feeling of actually being there.
Knowing where AI works — and where manual control matters more
A client from a strategic business development context wanted to test whether AI-generated video could accurately reproduce a specific branded headset from reference photos.
I ran multiple rounds across tools and prompt strategies. The AI consistently generated convincing atmospherics and visual mood — but could not reproduce the exact product with brand-level accuracy.
This is the kind of production decision that matters: knowing when AI delivers useful speed, when it needs tighter human control, and when manual craft is the better path. Both options have their place — the value is in making the right call.
Most AI agents are still being judged by the wrong standard.
The real question is not whether they can do something impressive in a demo. It is whether they reduce operational load without making the workflow harder to trust.
That is the lens behind these workflows: less friction, more reliability, and systems you can actually explain, audit, and keep under control.
Not mockups. Not demos. These run on my own server.
Every workflow shown here is live — built and maintained on a self-hosted n8n instance in Nuremberg, Germany. This is the same kind of setup I build for clients. Click any screenshot to see the full workflow.
Chatbots
8 sub-workflows
Crawls a WordPress site automatically via Sitemap and REST API, extracts clean text, deduplicates, uploads to Nextcloud, and feeds content chunks into self-hosted OpenMemory (mem0) — powering a truthful website chatbot that only knows what's actually on the site.
Part of a multi-workflow automated book-writing system. Pulls text chunks from Baserow, redacts sensitive content, sends to OpenAI for structured fact extraction, validates results, creates fact card and evidence records, handles retries and error states — all tracked in a self-hosted Baserow database.
Switch between the workflow and the resulting database table — the extraction logic and the structured output are both part of the system.
Operations
Published · live
Handles all authentication emails for a self-hosted operations platform via webhook — invite, password reset, access request, approval, rejection, and thread notifications. Routes each event type to the correct email template automatically. No manual step required.
These are my own workflows — not client work shown here for confidentiality. Client setups are built to the same standard.
One server. Everything under your control.
The core stack covers what most small teams need: file storage, database, automation, document editing, and delivery — all GDPR-oriented, all self-hosted in Germany.
Additional tools like identity management, monitoring, and uptime tracking are already installed on the server and can be activated depending on your needs. No extra setup time for these — they are ready to connect.
Everything is documented so you can maintain and extend it independently after handover.
These tools are set up and running on the server — no extra installation needed. Included where relevant for your use case.
From the field — on LinkedIn.
Where AI actually helps, where it breaks, where privacy-first infrastructure matters, and where communication gets stronger through better judgment instead of more hype.
Video AI nailed the vibe – and completely missed the product.
Where current video AI looks convincing at first glance but fails real-world product accuracy.
Read on LinkedInLess downtime. Faster ops. Your data, your rules.
Privacy-first infrastructure work: routing, cleaner operations, and setups that reduce noise instead of adding it.
Read on LinkedInAI is clever. It still doesn't feel when a shot finally lands.
Why manual editing judgment still matters — and why craft and automation should not be opposites.
Read on LinkedIn"If you stay silent, you'll stay broke."
On self-marketing, creative business realities, and why talent alone doesn't create traction.
Read on LinkedInI had ChatGPT build a patch for my app. Then I gave the same state to Claude for review.
What came back was not polish. It was a demolition of what had just looked like a solution — inside a real app-building process already 70 versions deep.
Read on LinkedInMost AI agents are still being judged by the wrong standard.
The real question is not whether they impress in a demo. It is whether they reduce operational load without making the workflow harder to trust.
Read on LinkedInAI agents are not risky because they are “smart.” They are risky because they have permissions.
Why tool use, exposure defaults, and boundary-setting matter more than shiny agent demos when real infrastructure is involved.
Read on LinkedInThe fastest AI automation pipeline is the one that gets you banned.
A blunt reminder that speed without legal and ethical boundaries can turn a content automation idea into operational risk instead of leverage.
Read on LinkedInTell me about your situation.
No pitch call first. Tell me what your business does, where the friction is, and what you want to improve. I read every message myself and respond within 24 hours, Monday to Friday.