Smart Contract Development for Algorand with TEAL and PyTeal
We develop smart contracts for Algorand using TEAL, PyTeal, and Beaker. Algorand is not EVM-compatible: AVM uses a stack-based model, and TEAL is an assembly-like language with strict limits. Program size is up to 8192 bytes, stack depth 1000, opcode budget per transaction only 20,000 (up to 320,000 in atomic groups). Every instruction matters: ed25519verify costs 1900 opcodes, keccak256 costs 130. Budget miscalculation can break a contract.
Algorand AVM offers advantages over Ethereum: instant finality (4.5 seconds vs 12 seconds), zero MEV, and fees under 0.001 ALGO per transaction—10 to 100 times cheaper than Ethereum layer 1. This suits DeFi, NFT marketplaces, and corporate solutions. However, the entry barrier is high: you need to grasp atomic groups, inner transactions, non-standard state model.
We accompany projects from analysis to deployment. Our portfolio includes token vesting, DAO voting, and AMM-like pools. Each project undergoes internal audit for vulnerabilities (reentrancy rare but possible via inner transactions) and opcode limit checks. We bring 5+ years blockchain development, 30+ projects on Algorand. The TEAL specification on GitHub confirms these limits.
AVM Specifics You Need Immediately
TEAL (Transaction Execution Approval Language) is stack-based for Algorand Smart Contracts (ASC1). Two program types: LogicSig (signs transactions without account) and Application (stateful with storage).
Key limits of AVM 10:
AVM limits
| Parameter | Limit |
|---|---|
| Program size (approval + clear) | 8192 bytes each |
| Stack depth | 1000 elements |
| Scratch space | 256 slots |
| Global state | 64 key-value pairs |
| Local state (per account) | 16 key-value pairs |
| Box storage (unlimited key-value) | pay per byte |
| Opcodes per transaction | 20,000 (base) |
| Opcodes with atomic group | up to 320,000 |
The 20,000 opcodes limit is not 20,000 lines. One keccak256 costs 130, ed25519verify 1900. Calculating budget is mandatory for complex contracts.
Why Atomic Groups Instead of Internal Calls?
Algorand contracts cannot call each other within one transaction. Instead, atomic groups—sets of transactions that all execute or all abort. Fundamental difference from EVM. For example, “flash loan → swap → repay” is implemented via a group of three Application Calls, verified using gtxn. Atomic groups provide composability without inner transaction overhead, but require correct index ordering.
How PyTeal Simplifies Development vs Raw TEAL?
PyTeal is a Python framework compiling to TEAL. It enables conditions, recursive loops, and ABI-compatible interfaces (ARC-4). For complex logic, PyTeal is indispensable. For simple LogicSigs or final optimization, raw TEAL works. Beaker (above PyTeal) adds auto-generation of ABI schemas and simplifies state management. We choose the right tool.
Tools and Stack
- AlgoKit—official CLI from Algorand Foundation. Project templates, local devnet via Docker (AlgoKit LocalNet), deployment, interaction.
- algopy (new Python framework)—compiles to TEAL via AVM compiler, static typing, strict ABI control. Preferred over PyTeal for new projects.
- algokit-client-generator—generates TypeScript client from ARC-4 JSON spec. Analogous to typechain for EVM.
- Algorand Sandbox / AlgoKit LocalNet—local node for development.
- Dappflow—web interface for inspecting transactions and contract state.
| Tool | Purpose | When to Choose |
|---|---|---|
| Raw TEAL | LogicSig, simple Approval, optimization | Minimal size, maximum control |
| PyTeal | Complex logic, ABI contracts | When conditions, loops, automatic compilation needed |
| Beaker | Quick start, auto-generation of ABI | New projects without legacy constraints |
We write tests in Python (pytest + algokit-utils) for unit and TypeScript (Jest + algosdk) for integration. Coverage checked via pytest-cov.
Contact us about contract architecture—we will help choose optimal type and tool.
How Development Looks
- Analysis (0.5–1 day). Determine contract type (stateful application vs LogicSig), state architecture (global/local/box), need for inner transactions, compatibility with ARC-4/ARC-20.
- Development (2–4 days). Write approval + clear programs. Parallel unit tests on AlgoKit LocalNet. Check opcode budget on critical paths.
- Integration. Generate ABI JSON, TypeScript/Python client. Test on Testnet before mainnet.
-
Deployment. Application deployed via ApplicationCreateTxn. Updatability via UpdateApplicationTxn if allowed. Immutability via hardcode
Int(0)in update/delete handlers.
Timeline: 3–5 days for medium complexity. Complex protocols: up to 2 weeks. Typical project cost ranges from $5,000 to $20,000 depending on complexity.
What's Included
- Security audit (reentrancy, underflow, access control) using Slither and manual analysis.
- Code documentation and ABI schemas.
- Test coverage (unit + integration) on Testnet.
- Communication at all stages.
- Support during mainnet deployment.
Typical Mistakes
Not accounting for minimum balance requirement. Each account opting in must have at least 0.1 ALGO + 0.025 ALGO per local state key. Box storage: 0.0025 ALGO per byte + 0.0025 per key. Forgetting breaks transactions with below min balance.
Mixing Application Call and Asset Transfer in wrong order. In atomic groups, transaction order matters. The contract reads gtxn 0—it must be exactly the expected transaction. Index confusion is a guaranteed bug.
Using LogicSig where an Application is needed. LogicSig signs transactions but does not store state. If logic requires global state (balance, counter, address list), a stateful Application is needed. LogicSig is for delegated authorization and escrow.
Proper opcode budget optimization can save up to 30% on fees for high-load operations—saving up to $3,000 per year for high-volume apps.
We help reduce development costs using ready-made templates and test automation. Contact us for project evaluation—get a consultation on contract architecture and cost estimation. Order Algorand smart contract development today.