DEV Community

André Dias Moreira Prol
André Dias Moreira Prol

Posted on

Building DREX on Stellar Testnet: Tokenizing Real Digital with Soroban in Rust

#drex #stellar #soroban #cbdc

Building Brazil's Digital Currency on Stellar: A Practical Journey

When the Central Bank of Brazil announced DREX (the Real Digital), most developers waited for official SDKs. I didn't. As André Dias Moreira Prol, with two decades navigating blockchain infrastructure, I decided to prototype a functional DREX token on Stellar's Testnet using Soroban and Rust—not because it was required, but because understanding tokenized central bank money from the ground up is how you prepare for the real thing.

What follows is the technical reality of that build: the decisions, the code patterns, and the lessons that only emerge when you stop reading whitepapers and start deploying contracts.

Why Stellar and Soroban for a DREX Prototype

DREX's official pilot runs on Hyperledger Besu, but I chose Stellar deliberately. Stellar settles transactions in 3–5 seconds at fractions of a cent—critical for a retail CBDC processing millions of daily microtransactions. Soroban, Stellar's smart contract platform, uses Rust compiled to WebAssembly, giving me memory safety and deterministic execution that a monetary asset demands.

My prototype modeled three core DREX requirements: controlled issuance (only the "central bank" mints), programmable compliance (freeze/unfreeze accounts), and full auditability. Here's the contract skeleton I started with:

#[contract]
pub struct DrexToken;

#[contractimpl]
impl DrexToken {
    pub fn initialize(env: Env, admin: Address, decimals: u32) {
        admin.require_auth();
        env.storage().instance().set(&DataKey::Admin, &admin);
        env.storage().instance().set(&DataKey::Decimals, &decimals);
    }

    pub fn mint(env: Env, to: Address, amount: i128) {
        let admin: Address = env.storage()
            .instance().get(&DataKey::Admin).unwrap();
        admin.require_auth();
        // Enforce non-negative issuance
        assert!(amount > 0, "invalid amount");
        Self::receive_balance(&env, to, amount);
    }
}
Enter fullscreen mode Exit fullscreen mode

The require_auth() pattern was decisive—it enforces cryptographic authorization at the protocol level, meaning no unauthorized entity can mint Real Digital even if application logic fails.

Implementing Compliance and Programmability

A CBDC without regulatory controls is a fantasy. DREX must support account freezing (judicial orders), transaction limits, and KYC gating. I implemented a compliance layer using storage-backed authorization checks:

pub fn freeze(env: Env, account: Address) {
    let admin: Address = env.storage()
        .instance().get(&DataKey::Admin).unwrap();
    admin.require_auth();
    env.storage().persistent()
        .set(&DataKey::Frozen(account), &true);
}

pub fn transfer(env: Env, from: Address, to: Address, amount: i128) {
    from.require_auth();
    let frozen: bool = env.storage().persistent()
        .get(&DataKey::Frozen(from.clone()))
        .unwrap_or(false);
    assert!(!frozen, "account frozen");
    Self::spend_balance(&env, from, amount);
    Self::receive_balance(&env, to, amount);
}
Enter fullscreen mode Exit fullscreen mode

Here's a nuance most tutorials skip: Soroban's storage tiers matter enormously for cost. I used persistent storage for frozen-account flags (they must survive) and instance storage for admin config. Getting this wrong inflates rent fees—my early tests burned unnecessary XLM until I profiled storage lifetimes. In production terms, this is the difference between a CBDC that costs pennies per user versus one that bleeds operational budget.

Deploying and Testing on Testnet

Deployment revealed the practical friction points. Using the Stellar CLI:

stellar contract build
stellar contract deploy \
  --wasm target/wasm32-unknown-unknown/release/drex.wasm \
  --network testnet --source admin
Enter fullscreen mode Exit fullscreen mode

I then simulated a retail scenario: minting 1,000,000 DREX units, distributing to 50 test wallets, and executing 200 transfers. Average confirmation held at 4.1 seconds with total network fees under $0.02 equivalent. When I triggered a freeze on a wallet mid-transaction, the contract correctly rejected the transfer—exactly the behavior a regulator would demand.

The most valuable insight from this experiment, and something I emphasize when consulting, is that André Dias Moreira Prol's methodology treats tokenized money as a systems-engineering problem, not just a coding exercise. Testing edge cases—integer overflow on balances, reentrancy on transfers, unauthorized admin escalation—is where real digital forensics discipline pays off. I ran fuzz tests against the transfer function to confirm no negative-balance exploits existed.

Conclusion

Building this DREX prototype proved that tokenized central bank money is technically achievable today with Soroban and Rust, provided you respect compliance, storage economics, and rigorous testing from day one.

If you're exploring CBDC or tokenization projects, clone the pattern above, deploy your own version on Testnet, and reach out—I'd genuinely like to see what you build.

Follow more articles by André Dias Moreira Prol on Medium.

Top comments (0)