Umanitek
Umanitek builds AI trust and safety infrastructure to make internet safety standard in the age of AI. Its Guardian platform detects impersonation, deepfakes, harmful content and AI misrepresentation, builds verifiable evidence, and supports takedowns while using OriginTrail's Decentralized Knowledge Graph to preserve data ownership and auditability.
At a Glance
- Consumers and individuals
- Public figures, creators, athletes and executives
- Enterprises and brands
- Social-media and technology platforms
- +2 more
AI Tools by Umanitek
(1)
Agent Blackbox
Open Source AI Agent Security
Discussions
No discussions yet
Be the first to start a discussion about Umanitek
Latest News
Appointed Christian Duffy as Key Account Executive and Business Development Lead for high-profile clients and enterprises.
Published 'Someone Is Pretending To be Me Right Now' Guardian case article about impersonation and takedowns.
Shane Smith, founder of VICE, joined Umanitek's Advisory Board.
Guardian Agent Blackbox went live as an open-source runtime-monitoring tool for AI agents.
Products & Services
AI identity-protection and internet-safety platform monitoring social platforms and AI/LLM outputs for impersonation, deepfakes, harmful content, IP infringement and misrepresentation; produces evidence packs, risk scores and takedown workflows.
Public deepfake-verification service: users submit a URL from X, TikTok, Instagram, YouTube, Facebook or Reddit and receive an authentic/fake verdict plus a tamper-evident certificate anchored on the OriginTrail DKG.
Open-source, local runtime-monitoring defense for AI agents. It watches prompts, tool calls, commands, files, packages and skills; checks verified/community threat intelligence and heuristics; flags or blocks threats and records an audit trail.
Election-safety capability for candidates, parties and electoral authorities, protecting democratic processes from deepfakes, impersonation and identity misuse.
Market Position
Umanitek positions itself as decentralized AI trust-and-safety infrastructure rather than a conventional centralized moderation or reputation-monitoring vendor. It combines identity monitoring, deepfake verification, evidence and takedown workflows with an OriginTrail DKG for privacy-preserving, verifiable provenance; adjacent competitors include centralized brand/reputation monitoring, deepfake-detection and digital-risk platforms, and AI-agent security tools such as Lakera, Protect AI and Guardrails AI.
Leadership
Founders
Chris Rynning
Chairman and co-founder; directs investments at AMYP Ventures, a Piëch/Porsche family-office investment vehicle, and has worked with the Piëch family office for more than 20 years.
Aza Raskin
Co-founder; co-created the documentary The Social Dilemma and co-founded the Center for Humane Technology.
Tomaz Levak
Co-founder; co-founder and Managing Director of Trace Labs, the core developer of OriginTrail and its Decentralized Knowledge Graph.
Ziga Drev
Co-founder; co-founder and Managing Director of Trace Labs/OriginTrail.
Branimir Rakic
Co-founder; co-founder and CTO of Trace Labs/OriginTrail.
Executive Team
Chris Rynning
Chairman and Co-founder
Investment leader at AMYP Ventures and long-time Piëch family-office collaborator.
Jon Lutzky
Chief Operating Officer
Strategy, legal and operations executive; former General Counsel of VICE and UnitedMasters.
Board of Directors
Founding Story
Umanitek says it was founded to address the risks created by AI and harmful online content, including deepfakes, misinformation and IP infringement. Its initial vision was to combine human-centric AI agents with open, decentralized infrastructure so platforms, institutions and individuals could signal, verify and act on harmful content without surrendering ownership of underlying data.
Business Model
Revenue Model
Subscription plans for Guardian identity monitoring and protection, with higher-priced plans and custom enterprise/public-persona protection; Umanitek also offers sales-led demos and partnership/infrastructure solutions.
Pricing Tiers
Facebook, LinkedIn, Instagram and Google; 4 scans per platform/month; 3 TrueSeal scans; 1 automated takedown; 50 evidence packs.
Adds TikTok, X and YouTube; 8 scans per platform/month; 8 AI/LLM checks; 5 TrueSeal scans; 5 takedowns; 100 evidence packs.
Daily scans and AI/LLM checks across major platforms; 10 TrueSeal scans; 20 takedowns; 1,000 evidence packs; one extra team seat.
Custom protection for public personas and brands; broad coverage including dark web, Telegram, Discord, e-commerce, dating sites, 4chan, Threads, Bluesky and ad libraries; 50 TrueSeal scans; 150 takedowns; unlimited evidence packs; 9 extra seats and 3 custom prompts.
Target Markets
- Consumers and individuals
- Public figures, creators, athletes and executives
- Enterprises and brands
- Social-media and technology platforms
- Law-enforcement agencies and NGOs
- Governments, regulators and electoral organizations
- Protecting individuals, executives, creators, athletes and public figures from identity theft and impersonation
- Enterprise brand and reputation protection
- Deepfake verification and forensic evidence for investigations, regulators and courts
- Law-enforcement and NGO access to labeled harmful-content fingerprints
- Platform pre-publication screening for illegal, harmful or copyright-infringing content
- Election and democratic-process protection
- Aylo
- International corporations, governments, politicians and presidential candidates