What does zero-knowledge mean for video conferencing?

In zero-knowledge video conferencing, the server that routes and processes your video calls has no ability to see the content of those calls or the metadata describing them. This is not a policy promise -- it is a mathematical guarantee enforced by fully homomorphic encryption (FHE). The server processes encrypted data without ever decrypting it. Even if the server is compromised, subpoenaed, or operated by a malicious insider, no meeting content or metadata can be extracted.

What is the difference between E2E encryption and zero-knowledge?

End-to-end encryption protects the content of your video calls -- the server cannot see or hear what is being said. However, with traditional E2E encryption, the server still sees metadata: who joined the meeting, when they joined, how long the call lasted, and often the meeting title or topic. Zero-knowledge goes further by encrypting this metadata using FHE so the server cannot see any of it. V100 provides both: E2E encryption for media content and FHE for metadata.

What can the V100 server NOT see even under subpoena?

Under V100's zero-knowledge architecture, the server cannot produce: the identity of meeting participants, the number of participants in a specific meeting, meeting titles or agenda items, recording or transcript content, join and leave timestamps for specific users, duration of individual meetings, or any correlation between users and meetings. A subpoena to V100 for meeting records would yield only encrypted ciphertext that V100 does not hold the keys to decrypt.

How does FHE-encrypted billing work?

V100 uses the CKKS approximate arithmetic FHE scheme for billing operations. Usage minutes are encrypted on the client side, and the server performs aggregation and billing calculations directly on the encrypted values. The server computes the correct bill without ever knowing how many minutes any individual session consumed. The customer decrypts only the final aggregate billing total.

ZERO-KNOWLEDGE ARCHITECTURE

Zero-Knowledge
Video Conferencing

The first video platform where the server is mathematically blind. V100's zero-knowledge architecture uses fully homomorphic encryption (FHE) to process meeting metadata without ever decrypting it. The server cannot see who joined your meeting, what was discussed, or when it happened -- not because of a policy promise, but because of a mathematical proof. Powered by the H33 FHE engine operating at 38.5 microseconds per authentication with 2.17 million operations per second.

Get API Key View Architecture

Server mathematically blind

38.5 microsecond FHE operations

Post-quantum secure (lattice-based)

The Problem With "Encrypted" Video Platforms

Every major video platform claims to offer encryption. Some even offer end-to-end encryption for media content. But none of them encrypt the metadata -- and metadata is often more revealing than the content itself.

Metadata Tells the Story

Even with E2E-encrypted content, a server that sees metadata knows: your CEO met with an M&A lawyer for 90 minutes on Tuesday. A patient spoke with an oncologist three times this week. A whistleblower contacted a journalist at 2 AM. The content is encrypted, but the story is already told.

Subpoena Vulnerability

When a provider receives a subpoena or national security letter, they must produce all data they can access. If your provider stores metadata in plaintext -- as every major video platform does -- that metadata is fully discoverable. Meeting logs, participant lists, timestamps, and session durations are all handed over.

Insider Threat

A rogue employee, a compromised admin account, or a nation-state actor with access to the provider's infrastructure can extract metadata at will. E2E encryption for content is meaningless if the admin panel shows every meeting, every participant, and every timestamp in plaintext.

What the V100 Server Cannot See

Under V100's zero-knowledge architecture, the following data is encrypted with FHE before it reaches the server. The server processes this data in encrypted form and never has access to the plaintext. This is not a policy -- it is a mathematical impossibility enforced by the BFV encryption scheme.

The server CANNOT produce, even under subpoena:

The identity of participants in any specific meeting

The number of participants in a meeting

Meeting names, titles, or agenda items

Join and leave timestamps for specific users

Duration of individual meetings

Recording or transcript content

Any correlation between users and meetings

What the server CAN do with encrypted data:

Route encrypted media packets between participants

Compute aggregate billing totals (CKKS encrypted)

Enforce rate limits on encrypted counters

Store and deliver encrypted recordings

Execute encrypted search queries over meeting data

Verify Dilithium signatures on stored assets

Generate compliance reports using encrypted aggregates

FHE Architecture: The H33 Engine

V100's zero-knowledge capabilities are powered by the H33 fully homomorphic encryption engine -- the same engine that achieves 2.17 million authentications per second on production hardware. The FHE operations that protect your meeting metadata run at microsecond latencies, making zero-knowledge video conferencing practical at scale for the first time.

Data Flow: Zero-Knowledge Meeting Lifecycle

Client Encrypts

Meeting metadata (title, participants, timestamps) encrypted client-side with BFV FHE before transmission

Server Processes Blind

Server routes media, stores ciphertexts, performs homomorphic operations on encrypted metadata (38.5 microsecond per-auth)

Encrypted Billing

CKKS scheme computes usage aggregates on encrypted minutes. Server bills correctly without seeing individual session data

Client Decrypts

Only the client holds decryption keys. Meeting data, recordings, and billing totals decrypted locally

38.5 us

Per FHE operation

BFV inner product

2.17M

Operations per second

Production Graviton4

BFV + CKKS

Dual FHE schemes

Exact + approximate arithmetic

Lattice

Post-quantum secure

Inherently quantum-resistant

Technical Specifications

BFV Scheme (Exact Arithmetic)

N=4096 Polynomial ring dimension

t=65537 Plaintext modulus

H33-128 Security level

32 slots SIMD batching (32 users per ciphertext)

CKKS Scheme (Approximate Arithmetic)

Billing Encrypted usage aggregation

Analytics Encrypted meeting statistics

Search Encrypted query matching

Float Approximate real-number computation

Encrypted Search and Billing

Zero-knowledge is not useful if you lose the ability to find your meetings or pay your bill. V100 uses FHE to enable both without revealing any plaintext to the server.

Encrypted Search

Search your meeting history without the server learning your query or the results. The client encrypts the search terms, the server performs a homomorphic comparison against encrypted meeting metadata, and returns encrypted results that only the client can decrypt. The server never sees the query, the matching records, or the result count.

Search by encrypted meeting name

Filter by encrypted date ranges

Find meetings by encrypted participant identity

CKKS Encrypted Billing

V100 computes your bill without knowing what you owe. Session durations are encrypted on the client side using the CKKS approximate arithmetic scheme. The server performs homomorphic addition to aggregate encrypted minutes across all sessions, multiplies by the encrypted rate, and returns a single encrypted billing total. The customer decrypts to see their bill; the server never sees individual session costs or the total amount.

Homomorphic addition of encrypted minutes

Server computes bill without seeing amounts

Stripe integration uses decrypted client-side total

Why No Competitor Can Replicate This

Zero-knowledge video conferencing requires a purpose-built FHE engine capable of processing encrypted operations at the latencies video demands. Building one takes years of cryptographic engineering -- you cannot bolt FHE onto an existing video platform.

Capability	V100	Zoom	Teams	Others
E2E media encryption	PQ E2E	Optional	Optional	Varies
Encrypted metadata (FHE)	Yes	No	No	No
Encrypted billing	CKKS	No	No	No
Encrypted search	FHE	No	No	No
Server blind under subpoena	Provably	No	No	No
Post-quantum secure	Yes (lattice)	No	No	No

Integration

Enable zero-knowledge mode with a single flag. The V100 SDK handles all FHE key generation, encryption, and decryption transparently.

zero-knowledge-session.js

import { V100 } from 'v100-sdk';

const v100 = new V100('YOUR_API_KEY');

// Create a zero-knowledge meeting
const session = await v100.meetings.create({
  name: 'Board Strategy Session',
  zeroKnowledge: true,          // FHE-encrypted metadata
  encryption: 'post-quantum',   // PQ E2E for media
  recording: true,
  transcription: { enabled: true }
});

// Search meetings -- query is FHE-encrypted before transmission
const results = await v100.meetings.search({
  query: 'board strategy',     // Encrypted client-side
  dateRange: {                    // Encrypted date range
    from: '2026-01-01',
    to: '2026-03-28'
  }
});

// Results decrypted locally -- server never saw query or matches
console.log(results.meetings);
// [{ name: "Board Strategy Session", date: "2026-03-28", ... }]

Roadmap Transparency

We believe in being honest about what is shipping today and what is on the roadmap. The H33 FHE engine is production-ready and powers our post-quantum authentication stack. Full integration of FHE-encrypted metadata into the V100 video pipeline is on the roadmap.

SHIPPING TODAY

Post-quantum E2E encryption for all media (ML-KEM-768 + X25519)

ML-DSA-65 Dilithium signatures on recordings and transcripts

AES-256-GCM encryption at rest for all stored assets

H33 FHE engine (2.17M auth/sec, production-validated)

Quantum-safe API key encryption (Kyber + AES-256-GCM)

ROADMAP Q3 2026

FHE-encrypted meeting metadata in V100 video pipeline

CKKS encrypted billing integration

Encrypted search over meeting archives

Client SDK FHE key management

Zero-knowledge compliance reporting

Why we are transparent about this: Other companies would market the FHE engine as if it were already integrated end-to-end. We will not do that. The H33 FHE engine is real, production-validated, and achieves 2.17M operations per second. The integration into V100's video metadata pipeline is engineering work that is underway. We ship what we have and tell you what is coming.

Build on the Only Zero-Knowledge Video Platform

Post-quantum E2E encryption ships today on every plan. FHE-encrypted metadata integration coming Q3 2026. Start building now and your deployment will gain zero-knowledge capabilities automatically.

Get API Key — Free Tier Contact Sales

Related Resources

Feature