> The ultimate Model Context Protocol (MCP 2.x) server to strip AI provenance tags (C2PA, EXIF, XMP, IPTC), disrupt invisible frequency watermarks, and simulate authentic camera sensor physics.
Are your AI-assisted images getting flagged, labeled, or algorithmically suppressed on LinkedIn, Twitter/X, Instagram, or stock marketplaces?
Modern generative AI tools (**Midjourney, DALL-E 3, Adobe Firefly, Stable Diffusion, Flux, Canva**) embed cryptographic C2PA / JUMBF provenance manifests, invisible spatial watermarks, and generation prompts directly inside image bytes. Modern social platforms automatically scan these tags to downrank organic reach or slap disclaimers on your content.
AI Image Metadata Cleaner gives creators, developers, and AI agents full sovereign control over their media.
🚀 Bypass AI Labels & Downranking: Prevent LinkedIn, Twitter/X, and social algorithms from penalizing or attaching "AI-generated" warning tags to your posts.
🔒 100% Prompt & Workflow Privacy: Strip proprietary generation prompts, negative prompts, seeds, model checkpoints, LoRA configurations, and ComfyUI workflow graphs embedded inside PNG/JPEG chunks.
📈 Maximum Organic Reach: Images without synthetic C2PA tags receive standard organic distribution instead of algorithmic throttling.
🎨 Eliminate the "Plastic AI Look": Inject realistic CMOS sensor photon noise and optical physics so images blend seamlessly with authentic photography.
🛡️ Pass Automated Authenticity Audits: Suspicious empty (0-byte) EXIF files get detected easily. This MCP embeds genuine camera hardware metadata (iPhone 15 Pro, Sony A7 IV, etc.) to pass heuristic checks.
🤖 Native Agentic Automation: Trigger automated sanitization directly from Claude Desktop, Antigravity, Cursor, or any custom MCP workflow.
Who It's For: Founders, Marketers & Growth Creators
How It Helps: Post professional infographics, concept banners, and visuals without triggering LinkedIn's reach-reducing AI badges or algorithmic dampening.
Who It's For: Social Media Managers, Meme Creators & Growth Teams
How It Helps: Publish visual assets across Twitter/X, Instagram, Pinterest, and Reddit with full organic distribution and zero platform penalties.
Who It's For: Freelancers, Executives & Remote Professionals
How It Helps: Remove the artificial "plastic" smoothness from AI headshots by adding realistic CMOS photon grain, optical falloff, and genuine camera metadata.
Who It's For: Digital Artists, Agencies & Prompt Engineers
How It Helps: Prevent clients, competitors, or web crawlers from extracting your exact Midjourney/ComfyUI prompts, seeds, and workflow graphs from downloaded images.
Who It's For: Design Studios & Full-Stack Developers
How It Helps: Automatically sanitize hundreds of images in bulk via multi-threaded parallel processing before client handoff or production deployment.
Who It's For: Digital Product Sellers (Gumroad, Etsy, UI Mart)
How It Helps: Ensure uploaded product mockups, wallpapers, and design templates have clean, authentic photographic metadata without flagged C2PA markers.
🧹 100% C2PA & Manifest Purge: Deep byte-level removal of C2PA JUMBF manifests (JPEG APP11, PNG c2pacaPt), EXIF, XMP packets, and generator tags.
🌊 Invisible Watermark Disruption: Sub-perceptual 0.3% micro-resampling safely disrupts frequency-domain watermarks (such as Google SynthID) without visual degradation.
📸 Camera Sensor Grain Simulation: Adds luminance-adaptive CMOS photon noise to eliminate telltale artificial smoothness.
🎯 Realistic Camera EXIF Injection: Simulates authentic shutter speed, aperture, ISO, and focal length from real camera profiles.
⚡ Full MCP Integration: Exposes 6 specialized tools natively callable by Antigravity, Claude Desktop, Cursor, and any MCP-compatible agent.
📁 High-Speed Batch Processing: Process entire directories in parallel with instant summary reports.
anti_ai_sanitize — [FLAGSHIP PIPELINE]Full end-to-end transformation in a single call:
Deep C2PA, EXIF, and XMP metadata purge
0.3% micro-resampling to break invisible frequency watermarks (SynthID)
Dynamic CMOS photon noise injection to remove synthetic smoothness
Realistic optical EXIF embedding from genuine camera hardware
strip_metadata_only — [LOSSLESS UTILITY]Instant zero-latency metadata stripping (C2PA, EXIF, XMP, IPTC, ICC)
Completely lossless — 100% pixel values are preserved untouched
inspect_image_metadata — [DEEP AUDIT]Scans raw image bytes for embedded C2PA JUMBF manifests and provenance tags
Detects generation prompts, seed parameters, and ComfyUI workflow graphs
Calculates an instant AI-detection risk score
inject_camera_profile — [STEALTH EXIF INJECTOR]Injects hyper-realistic camera hardware metadata into clean images
Simulates authentic focal length, ISO, aperture, lens profile, and camera model
batch_sanitize_directory — [BULK AUTOMATION]Recursively scans and sanitizes entire folders of images in parallel
Generates instant multi-file batch execution summary reports
get_camera_profiles — [PROFILE REGISTRY]Returns full technical & optical specifications of all supported camera rigs (iPhone 15 Pro, Sony A7 IV, Canon R6, Samsung S24 Ultra)
📱 Apple iPhone 15 Pro (iOS 17.5.1, 24mm f/1.78, ISO 80, 1/120s)
📷 Sony Alpha A7 IV (FE 24-70mm F2.8 GM II, f/2.8, 50mm, ISO 160)
📱 Samsung Galaxy S24 Ultra (6.3mm f/1.7, ISO 100)
📷 Canon EOS R6 Mark II (RF24-105mm F4 L IS USM, ISO 200)
git clone https://github.com/satyamkumar420/ai-image-meta-cleaner.gitcd ai-image-meta-cleaneruv venv .venvsource .venv/bin/activate # Or .venv\Scripts\activate on Windowsuv pip install -r requirements.txt