How to create unlimited viral videos, fully automatic & free!
Summary
This video demonstrates how to build a fully automated, local, and 100% free AI video-generation pipeline. Run entirely on a personal computer using open-source tools—including N8N for orchestration, Ollama for scriptwriting, Kokoro for text-to-speech, ComfyUI for image generation, and the NCA Toolkit for editing—this system behaves like a complete video production agency. Creators can utilize this workflow to post highly engaging short-form content, generate viral traffic, capture email leads, and earn substantial daily income through affiliate marketing or services without ongoing API costs or data privacy concerns.
Key Insights
The entire AI-driven video workflow runs locally, for free, and without internet dependency.
By using open-source tools that run directly on your own machine (such as N8N, Ollama, Kokoro, ComfyUI, and MinIO), creators can generate infinite video files. The entire backend works 100% offline, preserving absolute data privacy and eliminating the recurring cloud subscription fees or API billing common with tools like Kling, Runway, or ElevenLabs.
The 'Keyword Seeding' business model monetizes automated video traffic with minimal human effort.
Instead of relying on unstable ad revenue, this workflow generates highly targeted daily videos that prompt viewers to comment a specific keyword (e.g., 'hamster' or 'surf'). An automated secondary DM flow can then capture these leads into an email list, allowing solo operators to recommend digital products, online courses, or affiliate links for high ticket monetization.
The N8N automation acts as an entire video-editing department mapped onto a single canvas.
Rather than acting as a simple 'black box' generator, the N8N workflow models a physical creative agency. It breaks operations down into distinct color-coded roles: a copywriter for scripts, an announcer/audio lead for voiceovers, a creative director for scene descriptions, a graphic designer (via Stable Diffusion text-to-image), and a post-production editor for stitching elements together.
Sections
The Solo-Business Context & Traffic Problem
The missing piece for most online solopreneurs is consistent traffic generation.
While coaching thousands of students over a decade, the presenter noticed most people get hyper-focused on their core offer or service and completely neglect traffic creation. Posting high volumes of organic, short-form video is highly effective, but most struggle because manual video creation is exhausting and time-consuming.
Keyword seeding offers a path to rapid monetization and email list growth.
By offering a valuable freebie (e.g., a PDF guide or short course) in the call-to-action of short-form videos, creators can use automated comment tools to build lists of hot leads. This direct-to-list pathway is much easier to scale into $300 to $700+ daily revenue streams compared to relying on platform creator funds.
Traditional cloud-based AI video generators quickly become too expensive for volume posting.
While online SaaS video platforms offer simple start-up plans, their pricing tiers scale dramatically as soon as creators want to post 5 to 10 automated videos daily. Running local open-source setups bypasses these cost thresholds entirely.
Mapping N8N's Digital Video Agency
N8N acts as the master conductor orchestration tool running on your desktop.
N8N does not generate the content directly; instead, it serves as the project manager that coordinates tasks between local tools like Ollama, Baserow, and ComfyUI, passing data packets smoothly from one specialized node to the next.
The Copywriting Department is the first phase of the automated pipeline.
An Ollama AI agent acts as a short-form copywriter. Its sole job is to ingest the initial topic prompt and write spoken word directly for the camera, generating an optimized hook, a main script, and a direct organic call-to-action (CTA).
The Audio Department uses advanced speech synthesis to read the script.
The script is handed off to Kokoro, an open-source text-to-speech tool running on the user's computer. Kokoro generates realistic, human-like voiceovers using preconfigured vocal styles (such as the 'Bella' voice model) and automatically creates synchronized subtitle transcriptions.
The Creative Director maps out visual progression and image prompts.
Using math, the system automatically segments the voiceover audio length into roughly 5-second scenes. An AI agent acts as creative director, checking the spoken script of each specific segment to generate highly descriptive Stable Diffusion image prompts that remain relevant to the words being said.
The Graphic Design and Post-Production Editorial Departments assemble the final draft.
ComfyUI references the generated prompts to draw high-quality images. The NCA Toolkit then pulls these images, applies dynamic panning and zoom transitions to turn them into active video clips, layers the voiceovers and subtitles, adds background music, and exports the final file into a local database.
The Local Open Source Technical Stack
Ollama replaces costly OpenAI API requests for script writing.
Ollama operates as a local LLM node. It allows the system to remain highly modular as creators can freely hot-swap in the latest open-source text models (such as Llama, Qwen, or Mistral) to upgrade the scripting brain for free.
Baserow is used as a privacy-safe local database organizer.
Acting like a local, hyper-scalable Google Sheets, Baserow manages two tables: one table for overarching video details (scripts, vocal actors, status trackers, and target keywords) and a sub-table to organize scene data, corresponding prompts, raw images, and rendered video clips.
ComfyUI converts prompt text into high-fidelity image compositions locally.
ComfyUI utilizes local computer resources to generate art in styles like fantasy, cyberpunk, or anime. By exporting API JSONs from ComfyUI, users can quickly inject different workflows directly into their N8N loop to alter video styles instantly.
MinIO serves as a high-performance local S3 cloud for audio and media assets.
MinIO acts as an offline browser-accessible file system to cache royalty-free background music tracks, output media files, voice recordings, and direct links during intermediate rendering processes.
Demonstration: Training Syrian Hamsters Video
The presenter launches a Syrian hamster video generation test.
Using the initial form submission, the presenter inputs a Syrian hamster beginner tip topic, set to 25 seconds target length, fantasy art style, the 'Bella' voice model, and a custom keyword prompt of 'hamster' to fetch a free training course.
Ollama writes a targeted script with seamless visual framing.
Ollama instantly formats a script titled 'Tiny Paws, Big Training'. It writes natural spoken dialogue about positive reinforcement, treats, and user engagement, with a call-to-action matching the desired 'hamster' keyword.
Local graphics render is bounded mostly by GPU and system processing speeds.
While Apple M-series chips and dedicated Nvidia Windows cards drastically speed up rendering, users can run the models on a basic CPU or RAM. In this demo, each detailed Stable Diffusion image takes roughly 1 minute of render time.
The NCA Toolkit packages assets into a high-quality final video file.
The system processes 8 independent scene images, creates zoom effects, merges them with Kokoro voice files, burns stylized yellow subtitles across the lower frame, adds a thematic soundtrack, and outputs a complete, ready-to-publish short-form video.
Ask a Question
*Uses 1 Wisdom coin from your coin balance
