Synthetic media will be impossible to detect
Synthetic screenshots, tweets, cloned voices, and AI-generated posts can look almost identical to genuine content.
The trust layer for voice.images.video.
Make AI voice identifiable, compliant, and verifiable wherever it travels.
Built for EU AI Act · SB 942 · India IT Rules · C2PA
Incoming call00:03
Acme BankSupport agent
SessionBound to this call
Recovered fromThe audio itself
image.jpg2.1 MB
clip.mp40:12
Acme BankVerified by VerdaLLumen StudiosMark foundHHelio FilmsAll frames marked
One provenance layer across generated media. Sample data, fictional organisations.
When AI can make anything, origin matters more than ever
AI models can generate convincing text, image, audio, and video in seconds. That power is exciting and also makes it harder to know what is real.
Your brand is easy to spoof
Bad actors can copy your name, photo, and style.
Platforms sit in the middle
Social networks and media platforms face growing pressure to moderate harmful or misleading content without the tools to trace where it came from.
Trust erodes for everyone
As doubt spreads, audiences become sceptical even of authentic creators, journalists, and brands.
Built for the regulations already here.
Governments now mandate watermarks on AI media. Verda covers all media, AI-made or not.
In effect nowEnforceable or open for signature; existing EU systems by Dec 2
2027 → 2028Scope expands
SB 942
In effect · Aug 2, 2026California AI Transparency Act · California, USA
A hidden, permanent watermark in every AI-generated file.
Applies toGenAI providers with over 1,000,000 monthly users in California
RequiresA latent watermark and a free public detection tool
Penalty$5,000 per violation, counted again each day
EU AI Act
In effect · Aug 2, 2026Regulation (EU) 2024/1689, Article 50 · European Union
AI-generated media must be marked in a machine-readable way.
Applies toProviders and deployers in the EU market
RequiresMachine-readable marking, and disclosure of deepfakes and AI chat
PenaltyUp to €15,000,000 or 3% of worldwide turnover
India IT Rules
In effect · Feb 20, 2026IT (Intermediary Guidelines) Amendment Rules, 2026 · India
Synthetic content must be clearly labelled.
Applies toSocial platforms and the people who post on them
RequiresClear labels, with provenance metadata preserved
PenaltyLoss of Section 79 safe harbour
Code of Practice
Open for signature · Jun 10, 2026Code of Practice on Transparency of AI-generated Content · European Union
The EU code that defines how marking is assessed.
Applies toProviders and deployers of generative AI
RequiresTwo layers of machine-readable marking
If you don’t signYou prove adequacy yourself
AB 853
Phasing in · 2027 → 2028California AI Transparency Act (expansion) · California, USA
The California act extends to platforms, then to cameras and phones.
Applies toLarge online platforms and makers of capture devices
RequiresPlatforms show provenance from 2027; devices mark captures from 2028
Penalty$5,000 per violation
Verda Voice app
Verda screens the calls you don’t answer.
An app for your phone. Turn on call forwarding, and Verda picks up when you can’t.
A call you miss or reject is forwarded to Verda.
Verda’s voice agent speaks to the caller for you.
It checks for deepfakes and watermarks, then sorts the call.
If it’s urgent, you’re brought in with context.
Who it's for
The companies that generate AI media, and the people on the receiving end of it.
Unknown numbers, scam calls and cloned voices reach your phone every day.
- Answers the calls you miss
- Flags scam and spam
- Brings you in when it’s urgent
Your agents speak for you on every call. Caller ID can be spoofed, and AI audio now has to be marked.
- Identity bound to the voice
- Survives the phone network
- Meets AI audio marking rules
AI call centresAI sales and dialler platformsConversational voice AI
Explore AgentsThe law now requires what you generate to carry a durable, machine-readable mark.
- Marked at generation
- Recorded in a registry
- Free to verify, anywhere
Model providersContent generatorsEditing and post-production toolsAvatar and dubbing platforms
Explore EnterpriseAnyone can copy your name, your photo and your style.
- A verified profile
- V-marked posts
- Anyone can check a post is yours
Public figuresBrandsAgencies
Explore the Verda appYou see media every day with no way to tell where it came from.
- Check any image, video or audio
- See who made it and if it changed
- Free, no account
Distribution platformsResearchersAuthoritiesThe public
Try VerifyBuilt for the way you deploy AI
VoiceFor anyone who answers a phone
An app that screens the calls you don’t answer.
Explore VoiceAgentsFor voice agents and AI voice telephonyGive AI agents an identity people can verify.
Explore AgentsEnterpriseFor AI platforms and media infrastructureThe trust and compliance layer for image, audio and video.
Explore EnterpriseVerda appFor public figures, brands and agenciesA clear signal that a post is genuinely yours.
Explore the Verda appVerda Voice
Every AI call should be identifiable.
The identity is embedded in the speech itself, so it survives the phone network.
The call
An AI callerspeaking nowMarked as it speaks
Marked as the caller speaks
See the Verda Voice appimage.jpgMark found
MetadataIntact
OwnerLumen Studios
Mark read afterOriginal file
Images
Provenance people never see.
Embed provenance directly into generated imagery without changing what people see.
- OriginalMetadata intact
- CroppedMetadata stripped, mark found
- ScreenshotPixels re-captured, mark found
- The markImperceptible, in the pixels
Video
From generated frame to shared clip.
Persistent provenance for AI-generated and transformed video.
GeneratedVerda markEditExportShareVerified
Frame 0001, generated
Drop media, or pick a sample
Samples only. Check real files at verda.ai/verify.
Reading pixels…
Make trust travel with the media.
Voice, images and video. One trust layer.
bd@verda.ai