ChunavIQ: An Open, End-to-End Verifiable Remote-Voting Prototype for India's Legal and Identity Constraints, with Deniable Revoting, Hardware-Held Trustee Keys and Adversarial Dress Rehearsals
- 1Independent researcher
Abstract
India runs the world's largest elections: about 970 million electors in 543 parliamentary constituencies in 2024. Any remote-voting proposal for India must fit its election law, its identity infrastructure and its scale, and must say what it does not solve. We present ChunavIQ, an open-source research prototype that combines established techniques (threshold ElGamal, zero-knowledge ballot proofs, deniable revoting, Terelius-Wikström verifiable shuffles, designated-verifier proofs and an append-only Merkle log) into a complete system adapted to Indian constraints. Voters register a phone with either Aadhaar or PAN; since PAN checks only a name and birth date, the PAN path adds a one-time code to the mobile number on the electoral roll and a face match. A voter may vote again at any time, and a vote at a polling station on polling day cancels the online ballot; nobody, including a coercer, can tell which ballot counted. One national key is shared among trustees whose shares stay inside hardware security modules. We contribute (i) the adapted protocol and two independent implementations with two independent verifiers that cross-check each other's records; (ii) a capacity model for 1-1.5 billion electors built from measured unit costs; (iii) an open dress-rehearsal method that runs a constituency election over HTTP under seven kinds of attack and publishes every request; and (iv) a provision-by-provision analysis of the Indian legal changes such a system would need. Every record we produced verified. A first rehearsal forced the largest constituency's full design rate onto one machine and recorded only 72-99% of honest ballots per phase; after fixes identified from its published metrics, a rehearsal at 78% of that rate recorded every honest ballot, 97-100% of them within 30 s, while refusing every attack. ChunavIQ uses synthetic data only and is not proposed for binding elections.
Code: https://github.com/itsashiish/chunaviq (archived at https://doi.org/10.5281/zenodo.23442948). Demo: https://chunaviq.oglic.com
Related
- A MULTI-AGENT GAME-THEORETIC PROOF-OF-CONCEPT PROTOTYPE FOR MODELLING ADVERSARIAL FLASH LOAN EXPLOIT USING REINFORCEMENT LEARNING
- Radial–Angular Feature Shaping with an eXtra-Node and Prototype-Based Cosine Similarity Under Adversarial Perturbations
- No Action Without Proof: A Control Plane with Verifiable Constraints, Trust Aware Coordination, and Tamper Evident Accountability
- Constraints on Galactic Dark Photons from a DALI Prototype
- Statistical Guarantees for Covariance Estimation with Adversarial Corruption under Constraints
BibTeX
@misc{ramtekkar2026chunaviq,
title = {{ChunavIQ: An Open, End-to-End Verifiable Remote-Voting Prototype for India's Legal and Identity Constraints, with Deniable Revoting, Hardware-Held Trustee Keys and Adversarial Dress Rehearsals}},
author = {Ashish Ramtekkar},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23447533},
url = {https://zenodo.org/records/23447533},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - ChunavIQ: An Open, End-to-End Verifiable Remote-Voting Prototype for India's Legal and Identity Constraints, with Deniable Revoting, Hardware-Held Trustee Keys and Adversarial Dress Rehearsals
AU - Ashish Ramtekkar
PY - 2026
DA - 2026/10/18
PB - Zenodo
DO - 10.5281/zenodo.23447533
UR - https://zenodo.org/records/23447533
AB - India runs the world's largest elections: about 970 million electors in 543 parliamentary constituencies in 2024. Any remote-voting proposal for India must fit its election law, its identity infrastructure and its scale, and must say what it does not solve. We present ChunavIQ, an open-source research prototype that combines established techniques (threshold ElGamal, zero-knowledge ballot proofs, deniable revoting, Terelius-Wikström verifiable shuffles, designated-verifier proofs and an append-only Merkle log) into a complete system adapted to Indian constraints. Voters register a phone with either Aadhaar or PAN; since PAN checks only a name and birth date, the PAN path adds a one-time code to the mobile number on the electoral roll and a face match. A voter may vote again at any time, and a vote at a polling station on polling day cancels the online ballot; nobody, including a coercer, can tell which ballot counted. One national key is shared among trustees whose shares stay inside hardware security modules. We contribute (i) the adapted protocol and two independent implementations with two independent verifiers that cross-check each other's records; (ii) a capacity model for 1-1.5 billion electors built from measured unit costs; (iii) an open dress-rehearsal method that runs a constituency election over HTTP under seven kinds of attack and publishes every request; and (iv) a provision-by-provision analysis of the Indian legal changes such a system would need. Every record we produced verified. A first rehearsal forced the largest constituency's full design rate onto one machine and recorded only 72-99% of honest ballots per phase; after fixes identified from its published metrics, a rehearsal at 78% of that rate recorded every honest ballot, 97-100% of them within 30 s, while refusing every attack. ChunavIQ uses synthetic data only and is not proposed for binding elections. Code: https://github.com/itsashiish/chunaviq (archived at https://doi.org/10.5281/zenodo.23442948). Demo: https://chunaviq.oglic.com
ER -