We've been working on Minerva (https://zkesg.com/) — a ZK-STARK proof engine with a focus on making zero-knowledge proofs accessible without a cryptography background.
What it does:
The core use case is proving a claim about your private data without revealing the data itself. Classic ZK applications — but we've focused on making the proof generation entirely client-side via WebAssembly, so private inputs genuinely never touch a server.
Example: "My carbon emissions are below the threshold" → generates a proof → verifier gets pass/fail, nothing else.
The stack:
Winterfell (STARK prover by Polygon Miden) — FRI-based proofs, ~95-bit security
wasm-pack compiled to WebAssembly, running at near-native speed in the browser
Axum backend for proof storage and API key management
Next.js frontend with Juno (o4-mini-based AI oracle) for circuit generation from plain English
What's shipped:
Minerva — Prove truth. Reveal nothing.
Build, validate, and understand zero-knowledge proofs with AI-guided circuit design
10 circuit templates: age verification, income range, carbon compliance, credit score, supply chain, financial audit, KYC, voting eligibility, balance proofs, document authenticity
JS/TS SDK on npm: npm install @digitalforgestudios/minerva-sdk
Rust verifier SDK on crates.io: cargo add minerva-zkesg-sdk
Python SDK (PyO3/maturin) — macOS arm64 wheel available, PyPI coming
CLI verifier binary (minerva-verify) for offline proof verification
Public verify page at https://zkesg.com/verify — drag and drop any proof file
The circuit builder:
You can also use Juno (the AI oracle at /chat) to generate circuits from a description like "prove that a company's Scope 2 emissions are below 50,000 tCO2e" — it outputs a circuit spec and picks the right template. Useful for non-cryptographers who need custom constraints.
Minerva — Prove truth. Reveal nothing.
Build, validate, and understand zero-knowledge proofs with AI-guided circuit design
mcdoolz•1h ago
What it does: The core use case is proving a claim about your private data without revealing the data itself. Classic ZK applications — but we've focused on making the proof generation entirely client-side via WebAssembly, so private inputs genuinely never touch a server.
Example: "My carbon emissions are below the threshold" → generates a proof → verifier gets pass/fail, nothing else.
The stack: Winterfell (STARK prover by Polygon Miden) — FRI-based proofs, ~95-bit security wasm-pack compiled to WebAssembly, running at near-native speed in the browser Axum backend for proof storage and API key management Next.js frontend with Juno (o4-mini-based AI oracle) for circuit generation from plain English
What's shipped: Minerva — Prove truth. Reveal nothing. Build, validate, and understand zero-knowledge proofs with AI-guided circuit design 10 circuit templates: age verification, income range, carbon compliance, credit score, supply chain, financial audit, KYC, voting eligibility, balance proofs, document authenticity JS/TS SDK on npm: npm install @digitalforgestudios/minerva-sdk Rust verifier SDK on crates.io: cargo add minerva-zkesg-sdk Python SDK (PyO3/maturin) — macOS arm64 wheel available, PyPI coming CLI verifier binary (minerva-verify) for offline proof verification Public verify page at https://zkesg.com/verify — drag and drop any proof file
The circuit builder: You can also use Juno (the AI oracle at /chat) to generate circuits from a description like "prove that a company's Scope 2 emissions are below 50,000 tCO2e" — it outputs a circuit spec and picks the right template. Useful for non-cryptographers who need custom constraints.
Minerva — Prove truth. Reveal nothing. Build, validate, and understand zero-knowledge proofs with AI-guided circuit design
Verify page (no login): https://zkesg.com/verify npm SDK: https://www.npmjs.com/package/@digitalforgestudios/minerva-s... crates.io: https://crates.io/crates/minerva-zkesg-sdk Docs: https://zkesg.com/docs