MotionPrint Protocol Overview

MotionPrint is a cryptographic provenance protocol engineered for 3D motion interchange formats (glTF/GLB, BVH). Unlike server-centric approaches, MotionPrint implements a privacy-first, client-side architecture using WebAssembly.

The Attribution Problem

The scaling of 3D content creation via AI creates specific problems for existing attribution systems:

• Opaque Provenance - Difficulty tracing the training data lineage of AI-generated assets
• Registry Dependence - Licensing violations remain undetectable without access to central servers
• Format Erasure - Metadata is frequently lost during export or format conversion
• Privacy Risks - Verification systems that require file uploads expose proprietary data

Our Solution

MotionPrint addresses these challenges through:

1. Ed25519 Digital Signatures - Cryptographically verifiable authorship
2. Non-destructive Embedding - Provenance stored in existing metadata fields
3. Client-side Verification - Zero upload, zero tracking, zero account required
4. Format Agnostic - Works with GLB/glTF and BVH files

How It Works

Signing (Server-Side)

When you register an animation:

1. Generate payload with creator ID, license, timestamp, nonce
2. Canonicalize JSON for deterministic verification
3. Sign with Ed25519 private key
4. Inject signed payload into file metadata
5. Generate PDF certificate with QR code

Verification (Client-Side)

When you verify an animation:

1. Load file locally in browser memory
2. Extract signature payload from metadata
3. Verify Ed25519 signature against public key
4. Display verification result

The animation file never leaves your device.

Privacy Guarantees

MotionPrint is built on three non-negotiable privacy principles:

Principle	Implementation
Zero Upload	Animation file is never transmitted to any server
Zero Account	Verification requires no user registration
Zero Tracking	No cookies, fingerprints, or persistent identifiers

These guarantees are architectural, not policy-based.

Cryptographic Foundation

Why Ed25519?

Property	Ed25519	RSA-2048	ECDSA P-256
Signature Size	64 bytes	256 bytes	64 bytes
Public Key Size	32 bytes	256 bytes	64 bytes
Verify Rate	~75k/s	~15k/s	~20k/s
Deterministic	Yes	No	No

Ed25519 was selected for:

• Performance - 75,000+ verifications per second
• Compactness - 64-byte signatures fit in text headers
• Determinism - No RNG during signing eliminates vulnerabilities

Library

We use the audited @noble/ed25519 library for all cryptographic operations.

Supported Formats

GLB/glTF

Embedded in the asset.extras.motionprint field:

{
  "asset": {
    "version": "2.0",
    "extras": {
      "motionprint": {
        "payload": {
          "creatorId": "green",
          "licenseId": "restricted",
          "issuedAt": "2025-12-25T22:52:51.331Z",
          "method": "extras-v1",
          "nonce": "f707614d6a07ff66d51933d6eeb3e0b0"
        },
        "signature": "0xeb1f72a31d84eea35c2aa249..."
      }
    }
  }
}

BVH

Embedded as a Base64-encoded header line:

TRZ-WM: eyJwYXlsb2FkIjp7ImNyZWF0b3JJZCI6...
HIERARCHY
ROOT Hips
...

Performance

Verification completes in under 10ms on commodity hardware:

Operation	Mean (ms)	95th % (ms)
GLB Payload Extraction	1.2	2.1
BVH Header Parsing	0.4	0.8
Ed25519 Verification	3.8	5.2
Total Latency	5.4	7.5

File Size Overhead

Format	Overhead	Notes
GLB	~350 bytes	JSON chunk expansion
BVH	~250 bytes	Header line addition

Negligible impact on file size.

Next Steps

• Verification Guide — how to verify animations
• API Reference — integration documentation
• Free Verification Tool — try it now in your browser
• Blender Add-on — desktop integration for in-Blender verify
• Roadmap — Unity / Unreal plugins, batch verification, REST API on the way

---

Read the full technical paper on Zenodo. Last updated 2026-04-29.