The Context Engine & Variable Explorer
Every asset Mass generates is grounded in one deterministic fact-store about your offer. This is how that engine works — and the full catalog of 303 variables you can inject into any prompt.
16 categories · 303 variables · Updated for the current release
What the Context Engine is
A deterministic fact-store and resolver that turns your business context into structured, reusable data — then injects it into every AI prompt.
When you describe your offer during onboarding, Mass doesn't just hand your words to a model. It distills them into a structured OfferContext — a fact-store of positioning, audience, offers, copy, and brand voice. Every generation downstream (funnel pages, emails, ads, automations) reads from that same context, so your product name, price, and promise stay identical across every asset.
- Deterministic — the same context always resolves to the same values — no guessing per generation.
- Single source of truth — change the price or promise once and every funnel page, email, and ad updates with it.
- Dot-path addressed — every fact has a stable path like offer.main.price you reference with {{offer.main.price}}.
- Graceful fallbacks — missing facts resolve to safe placeholders instead of breaking the prompt.
How resolution works
You write {{variable.path}} in any prompt; the engine swaps it for the matching fact before the prompt reaches the model.
At generation time resolveTemplate scans the prompt for every {{...}} token, looks each dot-path up in the current campaign's context, and substitutes the resolved value. Missing facts fall back to safe placeholders rather than breaking the prompt.
- 1
Describe
You describe your offer during onboarding — or paste a URL, file, or raw text. Mass interviews you for anything still missing.
- 2
Distill
Your inputs are distilled into structured facts: positioning, audience, offers, copy, and brand voice.
- 3
Store
Those facts become the OfferContext fact-store — one deterministic source of truth for the whole system.
- 4
Resolve
At generation time, resolveTemplate swaps every {{variable.path}} for the matching fact before the prompt reaches the model.
Example
Template: Get {{offer.main.name}} for just {{offer.main.price}}
Context: offer.main.name = "The Launch System"
offer.main.price = "$497"
Resolved: Get The Launch System for just $497The fact-store layers
The context is organized into 16 addressable categories. Each one is a namespace of related facts you can reference by path.
Global instructions and persona that frame every AI generation.
Positioning, awareness stage, beliefs, and triggers distilled from your inputs.
Raw research signals — audience, market, and competitor findings.
Complete generated artifacts available for reuse and remixing.
Your main offer, order bumps, upsells, and pricing.
Headlines, hooks, bullets, testimonials, and punchlines.
Funnel structure, steps, and the flow between pages.
CTAs, guarantees, urgency, and trust-building elements.
Nurture and follow-up email content.
Social and ad copy tailored per platform.
VSL and video script blocks.
Curriculum, modules, and lesson content.
Persuasion levers — scarcity, authority, and reciprocity.
Tracking, KPIs, and measurement fields.
Brand colors, fonts, and visual styling.
FAQs and objection-busting responses.
Variable Explorer
Search, filter, and copy any of the 303 variables. Paste the syntax into any prompt field — or type {{ to trigger autocomplete in the builder.
Showing 303 variables