GhostSignal is a protocol for embedding and extracting structured data directly from media streams such as video and audio.
It enables media itself to carry data — independently of platforms, trackers, or external APIs.
A project by Patternity.
Media dominates the web, but it is effectively data-blind.
Today, attribution, analytics, coupons, and Web3 interactions rely on:
- redirects and tracking links
- platform-specific APIs
- external metadata detached from the content itself
GhostSignal introduces a native data layer inside the media stream.
- Proof-of-watch / proof-of-play
- Embedded campaign identifiers
- Coupons and claims carried by the content
- Privacy-first attribution
- Web3 minting and claims without links or QR codes
The data travels with the media — wherever it goes.
- Structured data is encoded into the media stream using a deterministic transport.
- The media is distributed normally (files, streams, platforms).
- A decoder (browser extension, SDK, or app) extracts the embedded data locally.
- Applications react to the extracted signal.
No platform integration required.
- Status: Draft (v0)
- Current transport: Visible Byte Band (VBB) for demonstration purposes
See the protocol specification:
A browser-based reference demo is available.
The demo focuses on the GhostSignal transport layer and demonstrates:
- embedding data into a video stream using the VBB transport
- exporting the resulting video
- extracting embedded data from the video
- real-time visualization of the decoding process
The demo does not yet implement the full GhostSignal v0 data layer (META bootstrapping, stream layouts, and raw payload framing), which are defined in the specification.
Demo repository:
👉 ghost-signal-demo
Possible next directions (not commitments):
- Additional transports (less visible / more robust)
- Error correction and resilience to transcoding
- Optional authenticity (hashing, signatures)
- Tooling: encoder/decoder SDKs and reference implementations
- Browser extensions and application integrations
GhostSignal is developed under Patternity — a research and engineering organization focused on hidden patterns, signals, and behavioral structures in interactive systems.
Apache License 2.0