The missing economic infrastructure for linguistic diversity in AI. African and underrepresented language communities earn perpetual, on-chain royalties every time their datasets power an AI model.
| Resource | URL |
|---|---|
| ๐ Web App (Live) | https://lingualayer.vercel.app |
| ๐ฆ Monorepo | https://github.com/linguaFoundation/lingualayer |
| ๐ Testnet Explorer | https://stellar.expert/explorer/testnet |
| ๐ DatasetRegistry Contract | CBET4YWSMIZB3LGLVTDKQJ5HXQAPQGM3NKGXJLJEJQNF7TBDOVMXUOK |
| ๐ฎ QualityOracle Contract | CCJVLNJ5O4NHIFMJMYZRFYIBRFM3WS7BKGYGWIQXNQYFXQTUYAEZQR5 |
| ๐ฐ DataCommission Contract | CDTGZ2PFUODWQFKLCMF2XZ7NY2HQPJFN3BQKUOIYRBKL5VWKRQBZLMJ |
| ๐ค๏ธ LicenseRouter Contract | CAQPZ5IK3WSHZTRQUAEKK3GMVKHSMWUFMJ4RFBOQ2QYQAUWK2TGZJMK |
| ๐ธ RoyaltySplitter Contract | CBQPFV7LQRSQOJTLXE7LIQXZSPJHBV7LQYSVYUUQZFJTJDZMUFKGQHJ |
LinguaLayer is a decentralized protocol built on Stellar Soroban that enables fair, transparent ownership and monetization of multilingual AI training datasets.
Every Yoruba sentence, every Swahili paragraph, every Igbo audio clip scraped from the web powers billion-dollar AI models โ yet the communities who speak those languages receive no compensation, attribution, or control. LinguaLayer changes this:
- On-chain attribution โ every dataset is registered with its contributors' Stellar addresses and share allocations
- Automatic royalties โ every license purchase triggers a programmable split to all contributors instantly
- Quality-gated multipliers โ certified curators attest dataset quality on-chain, boosting royalties for high-quality contributions
- Trustless commissioning โ AI companies post USDC bounties for specific language datasets; funds are held in escrow and released on delivery
Trusted language curators stake XLM and submit quality attestations (score: 0โ100) for any registered dataset. Scores aggregate into a quality tier that feeds a royalty multiplier applied on every license purchase:
| Tier | Score Range | Royalty Multiplier |
|---|---|---|
| ๐ Platinum | 90โ100 | 1.5ร |
| ๐ฅ Gold | 75โ89 | 1.25ร |
| ๐ฅ Silver | 50โ74 | 1.0ร |
| ๐ฅ Bronze | 0โ49 | 0.75ร |
- Curators who submit malicious or outlier attestations face slashing of their stake
- A minimum attestation quorum is required before a quality tier is assigned
- Quality scores are stored permanently on-chain and visible via the Dataset Detail page
AI companies and researchers can post USDC-denominated bounties for specific language datasets they need:
Post Commission โ USDC locked in escrow โ Contributor delivers โ Admin verifies โ Escrow released on-chain
- Commissions are browseable at
/bountiesโ the Language Bounty Board - Arbiter role available for disputed commissions
- Milestone-based escrow release for larger bounties
- Commission deadline countdown on every bounty card
We replaced custom Freighter-only connection logic with @creit.tech/stellar-wallets-kit, enabling native support for every major Stellar wallet:
| Wallet | Type |
|---|---|
| Freighter | Browser Extension |
| xBull | Browser Extension + PWA |
| Lobstr | Mobile + Extension |
| Hana | Browser Extension |
| Rabet | Browser Extension |
| WalletConnect | QR Code / Mobile |
| Ledger | Hardware Wallet |
| ALBEDO | Web-based Signer |
Replaced custom Sign-In With Stellar (SIWS) with the official Stellar SEP-0010 web authentication standard:
- Backend issues a signed challenge transaction using the
manage_dataoperation - Client signs with their wallet and returns the envelope
- Backend verifies the signature against the Stellar network
- Issues a JWT for authenticated API calls
- Duplicate registration prevention (hash-based deduplication)
- ISO 639-3 language code validation at the contract level
- Reputation score tracked per contributor address
- Weighted contributor share modification by dataset owner
- Admin upgrade path via WASM hash rotation
linguaFoundation/lingualayer/ (monorepo)
โ
โโโ apps/
โ โโโ web/ # Next.js 15 Frontend
โ โ โโโ app/ # App Router pages
โ โ โ โโโ page.tsx # Dataset Marketplace
โ โ โ โโโ /bounties/ # Language Bounty Board
โ โ โ โโโ /communities/ # Language Community Explorer
โ โ โ โโโ /governance/ # Community Governance Voting
โ โ โ โโโ /royalties/ # Contributor Royalty Dashboard
โ โ โ โโโ /licensing/ # License Purchase Flow
โ โ โ โโโ /roadmap/ # Public Roadmap
โ โ โโโ lib/
โ โ โโโ wallets-kit.ts # Stellar Wallets Kit integration
โ โ
โ โโโ backend/ # Fastify REST + WebSocket API
โ โโโ src/indexer/ # Soroban event indexer โ Postgres
โ โโโ src/routes/sep10/ # SEP-0010 auth endpoints
โ โโโ src/routes/datasets/ # Dataset listing + detail
โ โโโ src/routes/commissions/ # DataCommission API
โ โโโ src/routes/quality/ # QualityOracle attestation API
โ
โโโ contracts/ # Soroban smart contracts (Rust)
โ โโโ dataset-registry/ # Dataset attribution + contributor shares
โ โโโ license-router/ # License issuance + validation + revocation
โ โโโ royalty-splitter/ # Weighted royalty distribution (5% treasury)
โ โโโ quality-oracle/ # Curator staking + attestation + slashing
โ โโโ data-commission/ # USDC escrow bounty board
โ
โโโ docs/ # Protocol documentation
โ โโโ ARCHITECTURE.md # Contract interaction patterns
โ
โโโ .github/workflows/
โ โโโ ci.yml # Node App Build + Rust/Soroban Build CI
โ
โโโ README.md
All contracts are deployed and verified on Stellar Testnet.
| Contract | Address | Description |
|---|---|---|
| DatasetRegistry v3 | CBET4YWSMIZB3LGLVTDKQJ5HXQAPQGM3NKGXJLJEJQNF7TBDOVMXUOK |
Core dataset + contributor registry |
| LicenseRouter | CAQPZ5IK3WSHZTRQUAEKK3GMVKHSMWUFMJ4RFBOQ2QYQAUWK2TGZJMK |
On-chain license issuance & validation |
| RoyaltySplitter | CBQPFV7LQRSQOJTLXE7LIQXZSPJHBV7LQYSVYUUQZFJTJDZMUFKGQHJ |
Weighted revenue distribution |
| QualityOracle | CCJVLNJ5O4NHIFMJMYZRFYIBRFM3WS7BKGYGWIQXNQYFXQTUYAEZQR5 |
Curator staking + quality scoring |
| DataCommission | CDTGZ2PFUODWQFKLCMF2XZ7NY2HQPJFN3BQKUOIYRBKL5VWKRQBZLMJ |
USDC escrow + bounty board |
๐ All contracts are verifiable on Stellar Expert (Testnet).
- Node.js 20+
- Rust (see
rust-toolchain.toml) + thestellarCLI - A Stellar Testnet wallet (Freighter recommended)
git clone https://github.com/linguaFoundation/lingualayer.git
cd lingualayer/apps/web
npm install
cp .env.example .env.local
# Set NEXT_PUBLIC_STELLAR_NETWORK=testnet
npm run devApp runs at http://localhost:3000.
cd lingualayer/apps/backend
npm install
cp .env.example .env
# Configure DATABASE_URL, STELLAR_NETWORK=testnet, CONTRACT addresses
npm run devBackend runs at http://localhost:4000.
All 5 contracts must be built with the wasm32v1-none target (the current
soroban-sdk release no longer supports wasm32-unknown-unknown on Rust
1.82+ โ see rust-toolchain.toml), and license-router must be built
after quality-oracle, since its cross-contract client is generated from
quality-oracle's compiled .wasm file rather than a normal Cargo
dependency.
contracts/scripts/deploy.sh handles the build order, deploys all 5 with
the stellar CLI,
initializes them (creating and funding a deployer identity on testnet if
needed), and writes the resulting contract IDs to
contracts/deployments/<network>.env:
cd lingualayer
NETWORK=testnet SOURCE=my-deployer ./contracts/scripts/deploy.shTo build a single contract by hand instead:
cargo build --target wasm32v1-none --release -p dataset-registry
stellar contract deploy \
--wasm target/wasm32v1-none/release/dataset_registry.wasm \
--source-account <YOUR_IDENTITY_OR_SECRET_KEY> \
--network testnetUser connects wallet (Freighter / xBull / Lobstr / etc.)
โ
โผ
SEP-0010 Authentication (backend issues challenge โ wallet signs โ JWT issued)
โ
โผ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ Dataset Marketplace โ
โ Browse datasets โ filter by language โ view contributors โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโฌโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ
โโโโโโโโโโโโโโผโโโโโโโโโโโโโ
โ License Purchase โ
โ LicenseRouter.issue() โ
โโโโโโโโโโโโโโฌโโโโโโโโโโโโโ
โ
โโโโโโโโโโโโโโโโโโโผโโโโโโโโโโโโโโโโโโโ
โ QualityOracle โ
โ Reads quality score โ tier โ โ
โ applies royalty multiplier โ
โโโโโโโโโโโโโโโโโโโฌโโโโโโโโโโโโโโโโโโโ
โ
โโโโโโโโโโโโโโโโโโโผโโโโโโโโโโโโโโโโโโโ
โ RoyaltySplitter โ
โ Distributes USDC to all โ
โ contributor addresses by share % โ
โ 5% โ protocol treasury โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
| Phase | Status | Milestone |
|---|---|---|
| Phase 1 | โ Complete | DatasetRegistry + LicenseRouter + RoyaltySplitter deployed to Testnet |
| Phase 2 | โ Complete | QualityOracle + DataCommission deployed; multi-wallet support; SEP-0010 |
| Phase 3 | ๐ In Progress | Security audit of all 5 contracts |
| Phase 4 | ๐ Planned | Mainnet deployment + first 50 African language community onboarding |
| Phase 5 | ๐ Planned | AI company partnership program + data procurement API |
We welcome contributors! We have 154 open GitHub Issues labeled by component and priority:
[priority: critical]โ blocking mainnet launch[priority: high]โ important for v2 feature completeness[component: contracts]โ Soroban/Rust work[component: backend]โ Fastify API + indexer[component: frontend]โ Next.js UI/UX
See CONTRIBUTING.md for setup instructions and our PR workflow.
See SECURITY.md for our vulnerability disclosure policy.
Current security posture:
- All contracts have emergency pause mechanisms
- Admin upgrade path via WASM hash rotation (no proxy pattern)
- BytesN<32> metadata hash validated as non-zero before storage
- All
unwrap()calls audited and replaced with proper error handling in production paths - QualityOracle curator slashing prevents malicious attestations
Pre-mainnet: We will commission a full third-party security audit before deploying to Stellar Mainnet.
MIT ยฉ 2025โ2026 LinguaLayer Contributors