BetterBoxd

Problem Statement

BetterBoxd is a sovereign, trust-minimized film-rating network that brings cinema’s red-carpet prestige onto the blockchain. It solves three entrenched problems in movie discovery: (1) unverifiable “I-watched-it” claims, (2) review bombing and genre bias, and (3) closed, siloed rating data that cannot be composably reused.How it worksProof-of-Personhood loginUsers enter through aWorldcoin Mini-App. World ID nullifiers guarantee one human = one account, eliminating Sybil attacks from the outset.Watch verification via zkTLSA lightweight browser plug-in (or desktop companion) captures TLS handshakes and byte-range requests while the user streams a title on Netflix (Prime & Disney+ coming next).A zero-knowledge circuit proves“≥ 95 % of this specific title was streamed”without revealing any personal data.The proof andtitle_idare submitted to a verifier contract onWorld Chain; aProofVerifiedevent unlocks the rating UI for that user–title pair.Bias-corrected rating flowThe user selects a 0-10 score.An on-device Bayesian model adjusts for the viewer’s historical genre skew (“harsh on rom-coms, generous on sci-fi”), producing anAdjScore.A signed mutation writes just themovieIdandAdjScoreto a public Hypergraph dataset. Private taste vectors stay encrypted inPrivy.Open knowledge graphEntities:Movies,Genres,MovieGenres, andBetterBoxdRating(the live aggregate).Anyone—studios, data scientists, DeFi protocols—can query or index this graph; it’s published under an MIT-style data license.ENS-anchored identifiersEvery film is deterministically minted as an L2 ENS name (title.eth,title-2010.eth, etc.). These names serve as canonical URIs across dApps, while vanity usernames offer a future premium tier.Immersive front endReel scrollfor rapid discovery on mobile.A3-D node–edge graph(three.js) that visualises movies, users, and genre links in real time.“Watchlist Blend” recommendations surface titles liked by bias-similar peers.Tech stack (high-level)Smart contracts: Solidity (zkTLS verifier, rating relayer, ENS registrar) on World Chain rollup.ZK tooling: Noir + Halo2 for circuit generation; Risc0 prover for speed.Data: The Graph → Hypergraph (JSON-LD) for the public dataset; Privy for encrypted user graph.Front end: Next.js 14, React Native, Tailwind/ShadCN UI, three.js visualisation.Why it mattersVerifiabilityOnly viewers who provably watched a film can score it—first in the industry.FairnessBias-normalisation thwarts review bombs while preserving individual taste.ComposabilityOpen, on-chain ratings unlock new DeFi primitives (e.g., collateralised NFT film bonds) and analytics products for studios.Sybil resistanceWorld ID + ENS names give BetterBoxd a durable identity layer unavailable to Web2 competitors.Current status & demoRunning on testnet today:World ID sign-in, ENS mint, zkTLS proof production, on-chain verification, and Hypergraph rating write.Live demo: scan our QR code to rateInceptionand watch the global graph light up.BetterBoxd transforms movie love into tamper-proof, public truth—ready to premiere at EthGlobal Cannes.

Solution

Our submission is delivered as a World Mini App built with MiniKit; World ID handles Sybil resistance, while the integrated wallet and gas-free World Chain let users record their on-chain film engagement without paying fees. Because the entire flow—from proof-of-personhood to transaction relay—runs inside the Mini App, we qualify squarely for World’s “Best Mini App” bounty and, by extension, the World Pool Prize.Every motion-picture asset is automatically minted as an L2 primary ENS name (e.g., title.eth, title-2025.eth), establishing a canonical URI that any dApp can reference. Power users can also claim vanity ENS handles that link their on-chain identity to their cinematic footprint. Treating ENS not merely as a wallet alias but as the backbone of a public knowledge namespace speaks directly to the “Best Use of L2 Primary Names,” “Most Creative Use Case,” and overall “Best Use of ENS” prize tracks.Engagement submissions are emitted as contract events, indexed by a Subgraph, and immediately transformed into a publicly queryable Knowledge Graph via the grc-20-ts library. This pipeline demonstrates The Graph’s end-to-end data stack—from on-chain ingestion and Subgraph processing to tokenized knowledge representation and open API endpoints—making the project a strong fit for “Best Knowledge Graph App,” “Use of grc-20-ts Library,” and “Best Use of Subgraphs / Token API” awards.