Remix: fork what works, make it yours with AI
Remix takes a proven template, funnel, or winning asset and rebuilds it around your business — swapping the audience, offer, brand, and copy while keeping the structure intact. This guide covers the keep/change layers, the two remix modes, surgical document remixing, and where Remix shows up across the product.
9 min read · Updated for the current release
What Remix is
Remix is AI-powered forking: it clones a source asset into your workspace and rewrites the parts you choose, leaving the proven structure untouched.
The hard part of building a funnel isn't the layout — it's the structure that converts. Remix lets you keep that structure and change everything that should be yours. Point it at a marketplace template, a public funnel, or a winning showcase asset, pick what to swap, and Mass generates a new version under your account.
Because the skeleton is preserved and only the chosen layers are regenerated, you get a launch-ready starting point in seconds instead of a blank canvas.
The keep/change layers
A remix is defined by which layers you keep and which you change — audience, offer, brand, and copy.
Every remix decomposes the source into independent layers. You mark each one keep or change; Mass only regenerates the ones you flagged, so a brand swap doesn't rewrite your offer and an offer swap doesn't touch your voice.
- Audience — the target market and industry the copy and angle speak to.
- Offer — the product, pricing, and value proposition.
- Brand — name, voice, and visual identity.
- Copy framework — the angles and tone of the written copy, regenerated for stronger messaging.
Remix presets
Presets are one-tap layer combinations for the most common remix jobs.
- Different Niche — same funnel, new industry/audience (changes audience).
- Different Offer — same structure, new product/pricing (changes offer).
- Brand Swap — same everything, your brand identity (changes brand).
- Full Remix — keep the structure, change audience, offer, and brand together.
- Upgrade Copy — regenerate all copy with stronger angles (changes copy framework).
- Localize — adapt for a different market or language (audience + brand + copy).
Two ways to remix: prompt vs guided
Describe what you want in a sentence, or pick layers and fill in specifics — your choice.
- Prompt mode — type a freeform description of your business and the AI auto-detects which layers to change.
- Guided mode — explicitly toggle each layer keep/change and fill in the fields for the ones you're changing.
What you end up with
A remix produces a new onboarding context plus a cloned page/section skeleton, so the funnel canvas and launch checklist seed themselves from the cloned system — no follow-up setup required.
Surgical document remixing
When a template carries its real source markup, Remix edits the actual pages, emails, and images — not just a brief.
Marketplace templates published from an ingested system embed the full underlying blueprint: the original HTML, email, and image for each artifact. When you remix one of those, Mass runs the real blueprint engine over it and surgically edits each affected asset against your swap choices.
The result is saved as editable pages in your workspace (as drafts you can open immediately), so you get working copies of the real documents — already rewritten for your audience, offer, and brand.
Where Remix lives
Remix shows up anywhere there's something worth forking — the marketplace, funnels, community templates, and the showcase.
- Marketplace templates — remix a campaign/template you browsed or purchased into your own onboarding context.
- Funnels — any funnel marked template or public can be remixed; a remix count tracks how often it's been forked.
- Community builder templates — fork a shared builder project — files are re-scanned for safety on open, and the remix count increments.
- Showcase winners — every curated showcase item has a 'Remix this' surface that loads its recipe for editing.
Video Remix (reference-to-video)
Video Remix is a separate pipeline that analyzes a reference video and rebuilds it shot by shot.
Point Video Remix at a reference video and it extracts a structured recipe — a list of shots with subject, framing, camera move, background, and dialogue. Each shot is rendered independently (and can be retried/resumed on its own) by routing to the right model: talking-head shots to a text-to-video model with audio, b-roll and cinematic shots to others, product shots to an image-to-video model.
An optional audio bed is generated and the shots are stitched into a finished video. Because shots are denormalized, a single failed shot never forces you to re-render the whole thing.