Verivyx is the infrastructure layer that monetizes AI agent access to content — while keeping the web frictionless for real humans. Powered by Stellar and the X402 Protocol.
AI agents access and extract content at unprecedented scale — with zero compensation flowing back to the people who created it.
Bots crawl and extract content at industrial scale. No CAPTCHA stops them long-term. No payment flows to creators. The economics of the open web are broken.
AI agents have no interoperable protocol to autonomously discover and fulfill payment requirements. HTTP 402 exists but has never been implemented at scale.
CAPTCHAs frustrate real users while sophisticated bots solve them in milliseconds. Existing defenses are losing the arms race against AI-powered scrapers.
One unified layer that serves humans seamlessly, charges AI agents automatically, and pays creators instantly.
PoW + JA4 fingerprinting verifies real browsers silently in the background. Real users never see a CAPTCHA or delay.
HTTP 402 challenge → USDC micropayment on Stellar → instant content access. No human in the loop, ever.
Set your price, embed the 2KB script, and watch on-chain revenue accumulate. No manual withdrawals, no trust required.
Every request is automatically classified and routed — human, paying agent, or blocked bot. No per-request configuration needed.
Five purpose-built microservices. Go for performance-critical paths. Node.js for blockchain and ecosystem depth. No compromises.
Identity management, RBAC, API key issuance, Audit Logging, and analytics aggregation. Primary owner of PostgreSQL.
High-throughput X402 protocol implementation. Redis-cached paid sessions. Testnet ↔ Mainnet orchestration at the edge.
Edge gatekeeper. Extracts JA4 fingerprints, monitors PoW timing, decides human vs bot in real-time on every request.
The only service touching Stellar Horizon & Soroban RPC. Validates XDR, verifies signatures, submits on-chain.
Drop-in paywall.min.js. Web Worker PoW, browser fingerprinting, human/agent divergence. Zero framework dependency.
Each layer is independently effective. Defeating all five at once is economically and technically impractical.
| Layer | Technology | What it does |
|---|---|---|
L1 |
JA4 TLS Fingerprinting
Client Hello → hash
|
Identifies bot clusters even when rotating IPs. The fingerprint is tied to the underlying SSL library — impossible to spoof without changing the entire network stack. Detects coordinated bot farms sharing the same JA4 signature across thousands of IPs.
JA4 TLS 1.3
|
L2 |
Proof of Work
SHA-256
|
CPU-intensive cost that makes high-frequency mass scraping economically unviable. Humans naturally solve in 100–500ms. Any response under 50ms is flagged as a speed anomaly and classified as a bot — regardless of other signals.
PoW <50ms = bot
|
L3 |
X-Internal-Token Mesh
Zero-trust service auth
|
No microservice accepts a request from another without a shared INTERNAL_TOKEN header. External actors cannot call internal APIs regardless of network access. Service-to-service calls are verified at every hop.
Zero-Trust
|
L4 |
Soroban USDC Auth
On-chain multi-op TX
|
Forces AI agents to commit real economic stake per access. Every content request requires a verified on-chain USDC micropayment. No payment = no access. Financial cost makes automated mass extraction economically self-defeating.
USDC Soroban
|
L5 |
Idempotency Keys
UUID v4 · Redis dedup
|
UUID-based double-payment prevention stored in Redis. Agents can safely retry failed requests without risk of being double-charged. Every transaction has a unique key verified before settlement is accepted.
UUID v4 Redis
|
Automatic on-chain revenue split in a single atomic transaction. No trust required — mathematics enforces fairness.
Every tool chosen for performance, security, and production reliability at scale.
Join creators and developers building the next generation of AI-native content economics on Stellar.