Smart Parameter Help

The problem

AI models expose hundreds of technical parameters. Guidance scale, inference steps, LoRA weights, controlnet conditioning, denoising strength, temporal consistency, motion bucket ID — the terminology alone can feel overwhelming.

Most AI tools leave you guessing. You get a raw slider labeled guidance_scale with a range of 0–20 and no explanation. Or you’re pointed to API documentation written for machine learning engineers. As a video editor or motion designer, you shouldn’t need a machine learning degree to get great results.

Every control, explained

Every input field in modelBridge that isn’t self-explanatory has a ⓘ icon. Click it, and you get three things:

1. What it does — a plain-language explanation of what the parameter actually affects in your output
2. Value direction — what happens when you go lower or higher
3. Recommended starting point — a concrete number or range to begin with

Here’s what that looks like in practice:

Guidance Scale ⓘ “How closely output follows your prompt. 3–7 works best. Lower = freer, higher = more exact but stiffer.”

Strength ⓘ “How much the input changes. 0 = nearly identical, 1 = completely new. Sweet spot: 0.3–0.7.”

LoRA Scale ⓘ “How strongly the style filter applies. Start at 0.7. Lower = subtle, higher = more pronounced.”

Compare that to what most AI tools give you: a raw slider with a technical name and no context. That’s the difference.

The scale

This isn’t a handful of tooltips added as an afterthought. The parameter help system covers the full breadth of AI generation:

• Over 700 curated parameter explanations — every one written specifically for video editors and motion designers
• 1,000+ AI models supported — from video generation and image synthesis to voice cloning and audio production
• Zero generic fallbacks — every single ⓘ icon delivers real, specific guidance. No “see API docs” dead ends
• 1,026 total parameters classified — if a field doesn’t show ⓘ, it’s because we deliberately decided it doesn’t need one

Learn more links

Every help tooltip includes a “Learn more ↗” link that takes you directly to the relevant section in modelBridge Academy’s Parameter Reference — a comprehensive guide with over 100 parameter sections organized across categories including guidance, strength, LoRA, controlnet, audio, video processing, face animation, color grading, and more.

The Parameter Reference goes deeper than the tooltip: explanations use Premiere Pro and After Effects analogies, include recommended settings for different scenarios, and list common mistakes editors actually make.

Self-explanatory fields stay clean

Parameters like “Duration”, “Resolution”, and “Aspect Ratio” don’t show a ⓘ icon — because they don’t need one. We deliberately hide help on around 300 self-explanatory fields so the icon always means “there’s something useful here.” No noise. No clutter. Just guidance where it matters.

Why this matters

This system exists because we believe AI tools should guide you toward your desired result — not just expose raw controls and hope for the best.

Whether you’re trying your first AI model or your hundredth, the help system ensures you always know what a parameter does, what value to start with, and where to learn more. Parameters that seemed intimidating yesterday become creative tools today. You develop intuition for what guidance scale does, when to adjust strength, how LoRA weights change the output — all while working inside Premiere Pro.

The best creative tools don’t just give you controls — they help you master them.