Pulse monitors availability of AI agents, APIs, and services with ZK-proven random challenges, cryptographic signatures, and on-chain SLA proofs verified on zkVerify.
Real-time uptime data for all monitored agents and services. Every challenge is ZK-proven random, every response is cryptographically signed.
Five layers of verifiability, from random challenge generation to on-chain SLA proofs.
A ZK-proven random nonce is generated via Kurier's RNG circuit on zkVerify. The randomness itself is verifiable on-chain, preventing challenge prediction.
The challenge is delivered to the service endpoint. The service signs the nonce with its private key, creating a cryptographic receipt that Pulse cannot fabricate.
Pulse verifies the service's signature against its registered public key, records the latency, and stores the heartbeat with its status (success, timeout, or invalid).
Every hour, all heartbeats are aggregated into a Merkle tree. Each leaf contains the nonce, status, timestamp, latency, and RNG proof ID. The root commits the entire period.
A Groth16 proof is generated proving the SLA computation is correct: uptime percentage, total challenges, and whether the target was met. The proof is submitted to zkVerify.
What Pulse proves and what it doesn't. We believe transparency about limitations builds more trust than overclaiming.
Every proof is submitted to zkVerify and can be independently verified by anyone.
Monitor any service — AI agents, APIs, signal feeds, oracles. Every tier includes ZK-proven challenges and on-chain SLA proofs.