HyperDeteXHyperDeteXHyperDeteX
Detect AI-generated voices. Earn rewards by sharing your real voice.
Contribute Voice Data
Submit real & AI voice samples
Earn $DTX Tokens
Get rewarded for your contributions
Secure the Network
Help protect against voice fraud
The voice era is
under threat
Deepfakes, scams, and synthetic voices are evolving faster than detection.
Indistinguishable Voices
AI voices are becoming impossible to tell apart from human voices.
Rapid Spread of Fraud
Scams and deepfake audio are spreading at unprecedented rates.
No Verification
No easy way to verify whether a voice is authentic.
HyperDeteX | you
AI-Powered Detection
Advanced AI analyzes voices in real-time to detect synthetic audio.
Earn Rewards
Get $DTX tokens for contributing your voice to improve detection.
Community-Driven
Decentralized network ensures robust, unbiased verification.
| Soon
Detect AI-generated voices in real time, directly from your Telegram chat. Be the first to try it — follow our channels for the launch.
Join our Telegram channel
Get the bot link first, plus presale updates and project announcements.
@hyperdetex
Follow us on X
Real-time releases, partnership news, and community highlights.
@Hyperdetex
Four simple |
Contribute to voice AI detection and earn rewards in minutes
Upload your voice sample
MP3 or WAV format
Our model analyzes it
Returns a detection score
Valid samples are anonymized
Securely stored
You earn DTX tokens
For each accepted contribution
Tiered |
Our token-based reward system incentivizes quality contributions and long-term participation
How Tier Progression Works
Your reward tier is automatically determined by your DTX token holdings.
Token & |
Decentralized data collection with automated reward system via smart contracts
Token Distribution
150M
Total Supply
Token Details
Automated Reward System
Smart Contract Automation
Rewards distributed automatically via Base
Transparent & Secure
All transactions verifiable on-chain
Community-Driven
18.25% of tokens reserved for contributor rewards
Project |
Our strategic plan to revolutionize voice AI detection
Foundation
- ›Telegram bot launch + community demo
- ›v2 baseline audited: 100% fake / 99.48% real on 5 TTS vendors (in-distribution)
- ›Base Sepolia testnet deployment
- ›Smart contracts live: DTXToken, HoldRegistry, SampleRewardBridge
v3-Telco Foundations
- ›XLS-R 300M + AASIST + DANN model scaffold
- ›Channel-augmentation pipeline (replay IR + 8 kHz + G.711/AMR/Opus)
- ›Public datasets onboarded (ASVspoof 2019/2021, In-the-Wild, MLAAD)
- ›Bounty program beta — 20 contributors, 5k telco-grade samples
- ›Hold-to-Contribute gate & on-chain reward flow
Train & Benchmark
- ›v3-telco training: 5 phases (warm-up → DANN → calibration → distillation)
- ›Public leaderboard: EER, min t-DCF, FPR@FNR5%, ECE per dataset
- ›Milestone: EER < 3% on In-the-Wild benchmark
- ›RawNet3 edge model distilled (sub-300 ms on 2 s chunks)
- ›Partner pilot with Swisscom (data + case study in exchange)
Certification & Public Launch
- ›Fraunhofer AISEC certification engagement
- ›Bounty program public launch — target 100k samples
- ›arXiv paper submission + telco SDK v1
- ›AI-call opt-in flow (consent 3-silos, EIP-712)
Telegram App
Native mobile experience for voice contribution
Global API
Public voice verification API for developers