I’m building Sigillo, an API layer designed to combine STWO proof generation, SNIP-36 submission, and ZK compliance templates into single HTTP calls. Targeted at STRK20 applications that need audit-ready flows without running a prover or formatting Invoke V3 manually.
Current state (prototype, local only):
- Go API backend (/proofs, /verify, /submit, /audit) and STWO Rust bridge with SIMD backend, generating verifiable Circle STARK proofs locally (16KB at 64 rows, 52KB at 32768 rows)
- SNIP-36 formatter tested locally against starknet.go v0.18; Sepolia RPC confirmed SNIP-36-capable (spec v0.10.2). On-chain submission is grant milestone 2
- Seven gNARK circuits: four KYC patterns (age, sanctions, residency, income) plus three audit templates (selective disclosure, authorized view, selective audit with Merkle proofs) generating real verifiable proofs (1.3K to 43K constraints, 148ms to 4.1s)
- Sepolia account deployed (nonce 1, 200 STRK funded)
- SNIP-36 Phase 1 threat model reviewed by @ohadStarkWare
Context on scope: Atlantic (Herodotus) is a managed prover for Cairo programs via SHARP; public docs do not expose a starknet_proveTransaction-shaped endpoint, STWO, or compliance templates. Sigillo targets that specific gap.
What I am looking for now: 1 or 2 STRK20 teams (privacy pool, private DeFi, compliance flows) willing to do a 15 minute call to tell me what is real pain vs what I have assumed. Not recruiting customers, validating scope before applying for the Starknet Seed Grant.
DM on X (@gabrielrondon), on Telegram, or reply here.
Threat model: SNIP 36 Phase 1: threat model sketch · GitHub
Backend: GitHub - gabrielrondon/sigillo-backend: Zero-Knowledge Proof as a Service - Stripe for ZK Proofs · GitHub
Live: https://sigillo.tech