Anonymous Governance

Privacy-preserving voting dApp on Midnight — votes stay anonymous via ZK proofs. Same idea could power anonymous misconduct reporting. My first hackathon, built solo from zero blockchain experience.

Comment

Inspiration

This was my first hackathon and my first time building anything on blockchain. I wanted to explore what "privacy by design" actually means in practice, not just as a buzzword — so I built a governance/voting tool where votes are recorded on-chain, but no individual vote is tied to a voter's identity.

The idea that excites me most: the same privacy pattern used here for anonymous community voting could power something more sensitive — like confidential workplace misconduct reporting, where people need a verifiable way to be heard without fear of being identified.

What it does

Anonymous Governance lets a community:

  • Create a proposal
  • Vote yes or no
  • Close the proposal to reveal a final tallied result (Approved / Rejected / Tied)
  • View a history of past closed proposals and their outcomes

Individual votes are recorded through Midnight's zero-knowledge proof infrastructure, so the aggregate tally is public, but no voter identity is stored alongside any single vote in the governance ledger.

How I built it

  • Contract: Written in Compact, Midnight's smart contract language, based on the official Bulletin Board (Bboard) template from create-mn-app, extended with proposal voting and closure logic
  • Frontend: React + TypeScript + Vite, using MUI components
  • Wallet integration: Midnight's DApp Connector API
  • Network: Midnight preprod, with a locally run proof server (Docker) for zero-knowledge proof generation

I also extended the base template with:

  • A "Final Result" card showing the computed outcome (Approved/Rejected/Tied) when a proposal closes
  • A "Governance History" section so closed proposals aren't lost
  • A "Privacy by Design" explainer, written precisely against what the contract actually guarantees (not just marketing language)

Challenges I faced

The single biggest challenge was wallet connectivity. I initially built and tested everything using the Lace wallet extension, and hit a persistent bug where Lace's "Authorize" popup would appear but not register clicks at all — reproducible even in a completely fresh Incognito Chrome profile with a brand-new wallet. After extensive debugging (extension reloads, Chrome restarts, disabling other extensions, checking for Lace updates), I switched to the 1AM wallet instead, which resolved the issue immediately with no other changes needed.

As a complete beginner to blockchain development, I also had to learn, in real time: how Midnight's wallet connector API works, how zero-knowledge proof generation happens locally via a proof server, how DUST (the network's fee resource) is generated from NIGHT tokens, and how to read and reason about a Compact smart contract well enough to accurately describe its actual privacy guarantees (rather than overstating them).

What I learned

  • How zero-knowledge proofs are used in practice to separate "what happened" from "who did it"
  • That not all wallet extensions for the same network behave identically — switching wallets can matter as much as writing code
  • How to read a smart contract closely enough to know exactly what privacy guarantees it does and doesn't provide, and to document that honestly rather than overclaiming
  • That a working, honestly-scoped project is more valuable than a broken, overambitious one — I deliberately chose not to attempt adding one-vote-per-participant enforcement (which would require contract changes and a redeploy) this close to the deadline, and documented it as a known limitation instead

What's next

  • Add per-voter nullifier tracking to the voting circuits, so the contract enforces one vote per eligible participant while still preserving anonymity (true Sybil resistance, not just anonymous-looking votes)
  • Support multiple concurrent proposals instead of one active proposal at a time
  • Add deadline-based proposal closing instead of manual closing
  • Explore adapting the same privacy pattern for a dedicated anonymous misconduct-reporting tool

Built With

Share this project:

Updates

Submission history