content-provenance

Python implementation of C2PA: Coalition for Content Provenance and Authenticity.

Pure TypeScript implementation of C2PA manifest reading, validation, and creation

Trust metadata for AI agents — a stamp costs ~15 tokens, re-verifying costs 15,000. Agents stamp what they verify; the next agent runs 'akf check' and builds on it. pip install akf

PIN Architecture: a six-layer forensic system for detecting AI-generated and manipulated imagery. Fifteen independent analysis pins run concurrently: C2PA provenance, compression forensics, CLIP/SigLIP2/frequency detectors, Grad-CAM explainability, LLM adjudication and a calibrated XGBoost ensemble. Signed provenance outranks statistical inference.

Trust Identity Protocol (TIP) Specification v5.0. The open, post-quantum, federated standard for verified human identity and AI content provenance on the internet. Whitepaper at theailab.org/whitepaper, DOI 10.5281/zenodo.20722378.

An invisible watermark is still a signal. A lightweight CNN ensemble that detects SynthID — the imperceptible watermark in ChatGPT / GPT-Image-2 images — trained on a laptop and validated against the official verifier.

Cryptographic Proof of Effort (CPoE) — IETF Internet-Draft specification for verifiable attestation of human cognitive involvement in digital content creation, built on the RATS architecture (RFC 9334)

An ethical dataset supporting research in digital content provenance and authenticity, compliant with C2PA standards. Licensed for non-commercial use under CC BY-NC 4.0.

Local-first Unicode steganography detector, encrypted text watermark studio, and robust blind image watermark toolkit.

An innovative concept to safeguard visual media against generative AI. This repository is the main hub for the 'Adaptive Spider Web' project, detailing its layered defense, dynamic noise structure, and the verification app concept

MCP server for reading C2PA content provenance manifests from media files (Google Lyria AI MP3s, Adobe Content Credentials, etc.)

Industrial-grade, geometric-robust invisible image watermarking that survives rotation, crop, scale and regeneration. Open-source, CPU-only and multi-tenant.

Combat fake news with cryptographic image verification. Origin Lens analyzes C2PA Content Credentials and EXIF metadata to detect AI-generated content, verify digital signatures, and reveal complete edit history. Privacy-first open source iOS app with on-device verification. (arXiv:2602.03423)

An educational text-watermark specification, evidence-led cleaner and C2PA hygiene tool.

Image provenance for AI agents: the ChronoVerify MCP server. Verify a photo's capture time and provenance (C2PA validation, EXIF/XMP, pixel forensics). Tools: verify_image (typed verdict) and get_signed_report. Local via npx, or hosted at chronoverify.com/mcp/http.

Stdlib-first removal of AI provenance marks from text and files (Layer A text scrub, C2PA/EXIF/XMP file cleaners; MIT)

A free, open scale for declaring how a work was made with generative AI — six levels (0–5), CC0, readable by people and machines. Interoperates with C2PA, IPTC and W3C.

CPP (Capture Provenance Profile) - Open specification for cryptographic proof of media capture events. Features RFC 6962 Merkle trees for deletion detection, RFC 3161 timestamping, and optional ACE (Attested Capture Extension) for zero-knowledge biometric attestation. Part of the VAP Framework.

Label AI-generated images to comply with the EU AI Act (Art. 50): a visible watermark plus machine-readable EXIF metadata.

Sebuah konsep inovatif untuk menjaga media visual dari penyalahgunaan AI generatif. Repositori ini adalah hub utama untuk proyek 'Adaptive Spider Web', yang merinci pertahanan berlapis, struktur noise dinamis, dan konsep aplikasi verifikasinya