Motion Bridge docs

Everything from install to advanced workflows. Motion Bridge is your AI co-pilot inside After Effects — generate motion, write expressions, caption footage, score with SFX, and organise comps using natural language.

v1.2 · latest AE 2024+ · macOS & Windows Updated 3 days ago
Getting Started

Introduction

Motion Bridge is an AI co-pilot for After Effects. Tell it what you want in plain English — it understands your comp, writes the layers and expressions, and applies them instantly.

What Motion Bridge helps you do

  • Generate motion from a vibe prompt
  • Chat your edits with multi-turn iteration
  • Caption footage with Whisper in 12 styles
  • Score with SFX from Freesound, Pixabay, or local library
  • Organise comps — bulk rename, recolor, group
  • Vision mode — AI names layers by content

What you'll see in the panel

  • 5 tabs: Vibe, AI to AE, Caption, SFX, Organise
  • Chat history with iterative follow-ups
  • Model switcher — Claude, GPT, Gemini
  • Settings with API keys, brand palette, shortcuts

Installation

Motion Bridge ships as a signed .zxp for After Effects 2024+.

Quick install

  1. Download the installer. Get MotionBridge.zxp from your order receipt.
  2. Close After Effects. The installer needs AE closed.
  3. Run the installer. Double-click — ZXPInstaller signs and registers.
  4. Open AE. Window → Extensions → Motion Bridge.
macOS Gatekeeper warning? Right-click the installer and choose Open.

Manual install

If ZXP installer fails, use aescripts ZXP Installer or Anastasiy's Extension Manager.

System requirements

RequirementMinimum
After Effects2024 (24.0+)
OSmacOS 12+, Windows 10+
RAM16 GB
InternetRequired (BYOK)

Setup & API keys

Motion Bridge is BYOK — you pay providers directly. No markup, no proxy.

Supported providers

Anthropic (Claude)

Best reasoning. Get key.

OpenAI

Required for Cloud Whisper. Get key.

Google (Gemini)

Strongest vision. Get key.

Adding a key

  1. Open Settings. Gear icon, top-right.
  2. Paste your key. Each provider has its own field.
  3. Pick a default model. Sonnet 4 for chat, Haiku for cheap one-shots.
  4. Test connection. Click Verify key.
One key is enough to start. Add OpenAI later for Whisper, Gemini for Vision.

Your first prompt

Two minutes from panel-open to first generation.

  1. Open a comp. Motion Bridge reads dimensions and frame rate.
  2. Switch to AI to AE tab. Chat opens.
  3. Type a prompt. build a 3-card grid, brand red, stagger 0.1s, ease out cubic
  4. Press ⌘ Enter. Motion Bridge generates and executes.
  5. Iterate. "Make the cards smaller" → it adjusts.
⌘Z works. Every generation is a single undo step.
Guides

Writing effective prompts

The plugin is only as good as your prompt. These patterns get the most out of it.

Be specific about motion

Vague
make it fun

"Fun" has no motion vocabulary.

Concrete
bouncy entry, overshoot 15%,
0.4s, ease elastic

Specific easing, duration, amount.

Reference styles by name

Abstract
premium feel

Maps to nothing visual.

Visual
Apple keynote style — slow,
generous spacing, white on
warm grey

Named look the AI recognises.

Iterate in small steps

# Round 1
build a 4-column product grid
# Round 2
make them taller, 16:9
# Round 3
stagger left-to-right, 0.08s

Keyword vocabulary

WantUse these words
Snappypunchy, snap, overshoot
Slow elegancecinematic, linger, ease out cubic
Techscan, glitch, data viz
Organicwiggle, drift, breathe

Common workflows

Three end-to-end recipes — copy them, adapt them.

Recipe 1 — Social reel from scratch

  1. Vibe. vertical 9:16, Y2K, glossy → styled card.
  2. AI to AE. add a 3-shot product hero with stagger.
  3. Caption. VO audio → "TikTok-Beats" style.
  4. SFX. Search whoosh, drop on entries.
  5. Organise. Smart rename with Vision on.

Recipe 2 — Brand the project

  1. Open Brand settings. Paste hex codes, pick fonts, set voice.
  2. Toggle "Use brand colors" in Vibe.
  3. Save a brand template. Reuse across projects.

Recipe 3 — Bulk cleanup before handoff

  1. Organise → Bulk rename. Pattern {comp}_{n:02}.
  2. Color tag. Brand / Footage / Text / FX / Audio.
  3. Group orphans. Pre-comps loose layers.

Best practices

Pick the right model

TaskModelWhy
Multi-step editsSonnet 4Best reasoning
Quick generationsHaiku5× cheaper, fast
Look at footageGemini FlashCheapest vision
CaptionsWhisper-large-v3Best on accents

Save tokens with brand templates

If you reuse the same context every project, save it as a brand template. Motion Bridge prepends it to every prompt.

Cache your SFX

Downloaded SFX live in ~/MotionBridge/sfx-cache/. Point Motion Bridge at it as a "local library".

Troubleshooting

Panel won't open

  • Confirm AE 2024+
  • Re-run installer with AE closed
  • Enable Preferences → Scripting → Allow Scripts to Write Files and Access Network
  • Delete CEP cache: ~/Library/Caches/CSXS/

"API key invalid"

  • Re-paste — trailing space breaks it
  • Confirm billing on provider dashboard
  • Try a different model

Generations slow

  • Switch Sonnet → Haiku
  • Cloud Whisper is ~3× faster than Local

Captions misalign by ~1 frame

Toggle Frame snap in Caption settings.

Panels

Vibe panel

Generate styled motion cards from a prompt.

How it works

  1. Open the Vibe tab
  2. Type a vibe: cinematic, warm, slow
  3. Pick a motion preset (5 available)
  4. Click Generate

Motion presets

PresetFeel
DriftSlow horizontal, ambient
PopPunchy scale + overshoot
SlideSnappy directional entry
BreathIdle wiggle, organic
PulseBeat-synced scale

AI to AE

Chat with your timeline in plain English. Multi-turn.

Example prompts

# Layer creation
"Add a 4-column grid with 0.08s stagger"
"Build me 3 lower thirds for a podcast"

# Iteration
"Make it bigger"
"Switch to brand blue"

# Expressions
"Make this layer wiggle subtly"
"Sync opacity to audio amplitude"

Multi-turn behavior

Each message remembers the last 10. Use --explain at the end of any prompt for a plain-English breakdown.

Caption panel

Word-by-word captions with animated emphasis.

Workflow

  1. Select footage layer. Any audio-bearing video.
  2. Choose transcription. Local Whisper or Cloud.
  3. Pick a style. 12 presets included.
  4. Generate. Caption layers drop in, timed to the word.

Caption styles

Bold-Yellow, Karaoke, Minimal-Mono, TikTok-Beats, Subway, Beam, Podcast-Bold, Glitch, Hand-Written, Mono-Type, Comic, Classic.

Languages

Whisper supports 99 languages. Force language in Settings for code-mixed audio.

SFX panel

Search free SFX from 3 libraries, drop on the playhead.

Sources

  • Freesound.org — Creative Commons
  • Pixabay Audio — royalty-free, no attribution
  • Your local library — point at any folder

Searching

Type a description: whoosh, impact warm, ui pop. Filter by duration. Hover to preview, click to add at playhead.

Organise panel

Rename, recolor, group, trim across one comp or every comp.

Bulk rename rules

# Prefix
prefix: "BG_"
# Replace
replace: "v01""final"
# Number
pattern: "Card_{n:02}"

Color tagging

Preset color-codes matching layers. Scope: current comp or entire project.

Smart group

Pre-comps with sensible defaults — size matches selection bounds.

Vision mode

New in v1.2. AI looks at each layer's content and names it accordingly.

How it works

Samples a frame, sends to Gemini Flash, renames based on what's visible:

MVI_4382.MOV          →  person_speaking_closeup
logo_final.png        →  brand_logo_white
shot_06_v2.mp4        →  product_topdown_white_bg
audio_001.wav         →  voiceover_male

Cost

~$0.001 per layer with Gemini Flash. A 50-layer project = $0.05.

Reference

Brand settings

Configure once, every generation respects your brand.

FieldExample
Palette (5 hex)#FE3B3A #3334DE #0B842C #D07BF5 #111
Display fontClash Display
Body fontPublic Sans
Voiceplayful / neutral / formal
Custom rules"never stock photos"

Brand templates: save a complete config, switch per-project. Templates are local, no cloud.

Shortcuts

ActionShortcut
Send chat prompt Enter
Switch tab next ]
Switch tab prev [
Re-run last prompt R
Toggle vision mode Shift V
Open settings ,

Cost & usage

BYOK means you pay providers directly.

TaskModelCost
Simple layer creationHaiku$0.001
Multi-step editSonnet 4$0.01–0.05
Caption 1-min clipWhisper Cloud$0.006
Caption 60-min interviewWhisper Cloud$0.36
Vision rename per layerGemini Flash$0.001

Heavy week ≈ $10. Settings → Usage tracks spend with monthly budgets.

Privacy & data

What's sent to providers

  • Your prompt text
  • Comp metadata (layer names, dimensions, frame rate)
  • Audio file (Whisper transcription)
  • Sampled 720p JPEG (Vision mode only)

What stays local

  • Source footage and audio files
  • API keys (OS keychain / encrypted store)
  • Chat history, generation outputs, SFX cache

FAQ

Does Motion Bridge work with AE Free / trial?

Yes — any AE 2024+, including trial.

Per machine or per user?

Per user. One license activates on up to 2 machines.

Can I use my own SFX library?

Yes. Settings → SFX → point at any folder.

Are prompts used for training?

Prompts go directly to your provider. We don't proxy or store. Provider policies apply.

Refund policy?

14-day no-questions-asked refund.

Premiere version?

Yes — Video Bridge.

Updates

Changelog

What's new, version by version.

v1.2

Vision mode

3 days ago
  • New: Vision mode in Organise
  • New: Brand templates
  • Improved: Caption frame-snap toggle
  • Fixed: Memory leak on long chat sessions
v1.1

SFX panel + Gemini

2 months ago
  • New: SFX panel
  • New: Gemini support
  • Improved: Multi-turn history across model swaps
v1.0

Public launch

5 months ago
  • Vibe, AI to AE, Caption, Organise panels
  • Claude + GPT support
  • Whisper local + cloud

Roadmap

What's coming next.

v1.3

In progress

Target: next month
  • Voice-to-prompt — talk to the panel
  • Custom motion presets
  • Project-wide Brand templates
Later

Considering

Not committed
  • Team licenses
  • Plugin marketplace
  • DaVinci Resolve port

Still stuck?

We reply to every email within 24 hours.

Email Support →