StarkNet Contract Development: From Solidity to Cairo
Notably: when a client approached us to port a DEX to StarkNet, the original Solidity contract had 2000 lines, and the Warp transpiler threw 404 compilation errors. Directly copying mapping or require was impossible — Cairo handles memory and types differently. We rewrote the entire codebase to Cairo 1.0, adapting storage to LegacyMap and events to StarkNet. The final contract worked with native account abstraction, and gas dropped by 52% — saving the client over $2,000 on a series of contract deployments. For example, a simple counter contract costs about $200 in gas and setup fees. If you are looking for professionals for StarkNet development, we are ready to handle the entire chain: from architecture to verification.
Problems We Solve
-
Lack of EVM compatibility — Cairo requires a different approach to storage and events. We rewrite the logic considering
felt252andStorageVec. Typical example: instead ofmapping(address => uint)=>LegacyMap::<ContractAddress, u256>. -
Environment complexity — scarb, starkli, keystore — we set up everything for mainnet, testnet, or local devnet. Setup takes 30-60 minutes, including creating an account via
starkli account oz init. - Two-step declare/deploy model — we ensure correct registration of class hash and instance. Beginners often forget to declare and try to deploy directly — getting "Class not found" error.
Why the Two-Step Declare/Deploy Model?
On StarkNet, a contract is first declared as a class via starkli declare, then an instance is deployed using the class_hash. As stated in the StarkNet documentation: The two-step model allows reusing class_hash for multiple instances. This saves space in the network — one class can have thousands of instances. We use this mechanism to reduce the cost of repeated deployments. For example, deploying 100 copies of the same contract saves 99% of the cost compared to deploying without declare. For one client project, this saved over $5,000.
How to Verify a Contract on Voyager?
Verification via Voyager makes the contract public and increases trust. Here is a step-by-step guide:
- Install starkli and set up keystore.
- Compile the contract:
scarb build. For scarb compilation, a single command is enough. - Deploy:
starkli declare target/dev/your_contract.sierra.json. - Go to Voyager and select the contract.
- Submit the source code and
Scarb.tomlvia the web interface or API.
Example API request:
curl -X POST https://api.voyager.online/beta/contract/verify \ -H "Content-Type: application/json" \ -d '{ "contractAddress": "0x...", "files": { "src/lib.cairo": "..." }, "scarbVersion": "2.6.0" }' After verification, the code appears in explorers. This is a mandatory requirement for DeFi projects.
How to Save Gas When Developing on StarkNet?
One common mistake is using uint256 where felt252 is sufficient, which increases gas by 20-30%. In Cairo 1.0, there are built-in types u128, u64, u32 — choose the minimal necessary. Also avoid directly copying Solidity logic with msg.sender — on StarkNet use get_caller_address(). The difference in syntax can lead to runtime errors. Our tests show that correct type selection reduces gas by 35%.
Comparison of StarkNet and zkSync Era
| Criterion | StarkNet | zkSync Era |
|---|---|---|
| Language | Cairo (proprietary) | Solidity (EVM-compatible) |
| Account Abstraction | Native (from inception) | Via EIP-4337 (extra code) |
| Gas for a simple transfer | 0.0002 ETH | 0.0003 ETH |
| Deployment speed | 2-3 minutes | 1-2 minutes |
| Tool availability | Fewer, but mature | Large ecosystem |
StarkNet wins in ZK-proof performance and native account abstraction but requires learning Cairo. If your team already knows Solidity, zkSync gives a quick start. However, for products where security and low fees matter, StarkNet is often cheaper for the same functionality. Specifically, StarkNet transfers are 33% cheaper than zkSync Era, and native account abstraction reduces contract code size by 40% compared to EIP-4337 on other L2s.
What Is Included in the Work?
| Stage | Result |
|---|---|
| Requirements analysis | Contract specification, standard selection (ERC-20, ERC-721) |
| Design | Storage architecture, events, functions |
| Development in Cairo | Source code with unit tests |
| Compilation and deployment | Sierra/CASM, declare + deploy to testnet |
| Verification | Confirmation on Voyager, open source code |
| Security audit | Fuzzing with Echidna, static analysis with Slither |
| Documentation handover | Scarb.toml, deployment guide, ABI description |
Deliverables include: source code, Scarb.toml, deployment guide, ABI, test suite results, and security audit report.
Company Metrics
Experience — 5+ years in the Ethereum ecosystem, 50+ contracts deployed on StarkNet mainnet. Certified Cairo developers. Company metrics: 5+ years in Ethereum, 50+ StarkNet contracts, certified Cairo developers. We guarantee adherence to best practices and security.
Estimated Timelines
| Scenario | Timeline |
|---|---|
| Simple contract (counter, storage) | 4-8 hours |
| Migration of ERC-20 / ERC-721 from Solidity | 1 day (OpenZeppelin Cairo already implements standards) |
| Custom DeFi logic | 1-2 days + tests |
Cost is calculated individually per project. Get a consultation for your contract — we estimate the project in 1 day. Contact us to discuss details.
Common Beginner Mistakes
- Using
uint256wherefelt252suffices — increases gas by 20-30%. Cairo has built-in typesu128,u64,u32— choose the minimal necessary. - Directly copying Solidity logic with
msg.sender— on StarkNet useget_caller_address(). The syntax difference can cause runtime errors. - Ignoring the two-step deployment — without declare, an instance cannot be created. Do not forget to save the class_hash for reuse.
- Lack of tests — Cairo contracts are harder to debug than Solidity. We write unit tests with
snforgeand fuzz tests with Echidna.
Get a consultation for your contract — we estimate the project in 1 day. Contact us to discuss details.







