Skip to main content
About

Local-first by design

A faceless YouTube channel today needs a script tool, a voice tool, a captions tool, a music tool, a thumbnail tool, an SEO tool, a scheduler, and a cloud render farm. That's eight monthly bills before a single video ships. Phantomline replaces that stack with one local install.

What Phantomline is

Phantomline is a local-first AI video studio. It generates scripts, narration, captions, music beds, visual layers, MP4 exports, and YouTube publishing drafts, all on your own machine. No per-render API fees. No usage caps you have to count against your project margin. License keys validate offline. The only thing on our servers is the marketing site.

What runs locally

On the desktop install, the heavy AI work happens on your hardware:

  • Ollama handles the language model: Llama 3.1 by default, but any model you can fit in your VRAM. Scripts, titles, descriptions, hooks.
  • Kokoro TTS handles narration. 16 voices ship with it. The model is ~330 MB on first run, then everything's local.
  • MusicGen generates ambient backing tracks via HuggingFace transformers. Crossfade-loops to any length you need.
  • ffmpeg assembles the final MP4. Industry standard, no Adobe license required.
  • Forge / AUTOMATIC1111 (optional) for AI-generated scenes if you have it running locally on port 7861.

On the browser version (phantomline.xyz/app), the same pipeline runs in-browser via WebLLM (Llama 3.2 1B over WebGPU), Web Speech API for TTS, Web Audio for music, ffmpeg.wasm for MP4. After the first load, the whole thing is airplane-mode safe.

Why local-first matters

The math on cloud-only AI video tools breaks at volume. Imagine you're shipping 30 videos a month, which is pretty normal for a faceless channel that wants to grow. At 4 cents a render across the script, voice, music, and caption pipeline (which is honestly cheap for a cloud stack), you're paying $1.20 per video, $36/month, on top of the subscriptions.

Now scale that to multi-channel. Five faceless niches, 30 videos each, and the API meter is $180/month. Pure overhead before anyone watches a video.

Local AI inverts this. Once you've installed the models, every render is free. The hardware you already own does the work.

Why faceless YouTube specifically

The faceless niche has a unique workflow profile. Creators ship volume. Reddit story channels can publish daily, listicle channels often post 3x a day. The bottleneck is rarely creative; it's pipeline cost and tool friction.

General-purpose AI video tools (Pictory, InVideo, etc.) are designed for marketers making one explainer video a quarter. They're priced and shaped for that use case. Faceless creators get squeezed by per-export limits, watermark removal upcharges, and renew-or-die subscriptions.

Phantomline is shaped for the volume use case from the start: render queues, schedule drafts, project bundles, channel-aware analytics, vidIQ-aware Optimize Library for republishing old uploads.

What we believe

  1. Software you paid for should keep working when the company's website goes offline.
  2. If a feature doesn't need a server, it shouldn't have one.
  3. Pricing should be visible without a sales call.
  4. Lifetime tiers are a real promise, not a launch gimmick. Founding Lifetime customers get every Creator Pro feature forever, even if our pricing rises.

Who's behind this

Phantomline is co-founded by Kyle Anson McCool (co-founder of Makko AI) with Cesar Martinez and Zac Dillard — the same trio behind ProLocalBuilder. kyle@makko.ai for direct contact. Cesar handles UI work. The repo lives at github.com/Kyle-Anson-McCool/phantomline for the people who care.

Get started

Open the studio or see pricing. The free tier needs no card.