Live

Verifiable Video Manifest Protocol

Open source provenance and trust layer for AI generated and AI assisted video, built on top of C2PA. Answers questions about video origin, sources, prompts, models, edits, and guardrails through cryptographically verifiable manifests.

Live

Open source provenance protocol / AI generated video trust layer

Verifiable Video Manifest Protocol cover

What this proves

I can design and ship open infrastructure at the protocol level, not just applications, in a fast moving space where standards are still being defined.

Why this mattered

As AI generated video becomes indistinguishable from real footage, the field lacks a rigorous, open standard for proving what a video is, where it came from, what model made it, what edits were applied, and what guardrails were in place.

Why it mattered commercially

Media platforms, AI video companies, and enterprise content teams will need provenance standards as AI generated video becomes ubiquitous. Building open infrastructure now means being the reference implementation before the standards ossify.

Current status

Active public open source work. Repository contents and Apache 2.0 license were verified on August 30, 2026; adoption claims are intentionally limited to what the public repository proves.

What Gordon built and maintains

Designed the VVMP trust schema and cryptographic manifest structure on top of C2PA to cover the full lifecycle of AI generated video.

Built out the modular package architecture: trust schema, capture SDK, C2PA adapter, registry client, trust viewer app, and CLI.

Published the repo publicly under Apache 2.0 and structured it for external contribution and consumption via npm packages.

What was hard

C2PA was designed primarily for images and documents, adapting it cleanly for the temporal and generative specific properties of AI video required meaningful extension work.

Designing for modularity while keeping the manifest schema coherent requires real architecture discipline upfront.

The protocol has to be rigorous enough to be cryptographically meaningful but practical enough for real production pipelines to adopt.

Proof at a glance

C2PA

Built on open standard

Extends the Coalition for Content Provenance and Authenticity standard for AI generated video.

Apache 2.0

Open source

Fully public, open source, and structured for npm package publishing and external adoption.

View evidence →

6+ modules

Modular architecture

Trust schema, capture SDK, C2PA adapter, registry client, trust viewer, and CLI.

Created an open, cryptographically sound foundation for AI video provenance that teams can build on now.
Modular architecture means adoption can start at the schema or SDK level without a full platform rebuild.
Connects the gap between AI video generation and the emerging content authenticity ecosystem.

Sources and verification

Where the public evidence lives.

Project record reviewed August 30, 2026

Source code / GitHub

Verifiable Video Manifest Protocol

The public schema, SDKs, CLI, tests, C2PA integration, license, and development history.

Link verified August 30, 2026

Open evidence →