Oh My Sol vs Código for Solana development

Short answer

Código is a browser IDE for Solana with its own Solana-trained model, templates for Anchor and native programs, a test validator in every workspace and one-click deploy. Oh My Sol is a terminal harness that runs on your machine with the model you already use, adds Solana references, error decoding and toolchain checks, and closes a task only when your requirements pass on a fresh local Surfpool chain.

Who it’s for
Solana developers deciding between a cloud IDE and a local terminal harness for AI-assisted program work.
Last checked
, against Oh My Sol 0.1.0

What is the difference between Oh My Sol and Código?

Both are built for Solana, but they sit in different places. Código is a cloud platform you use in a browser. Oh My Sol (OMS) is a terminal app that runs on your own machine. Facts about Código below come from codigo.ai and its documentation.

CódigoOh My Sol 0.1.0
Where it runsBrowser (cloud IDE)Your terminal on macOS or Linux, or Windows through WSL 2
ModelIts own custom-trained model for SolanaThe model you connect: OpenAI, Anthropic, Xiaomi MiMo or OpenCode, by API key or subscription
Starting pointTemplates for Anchor and native programs, or a Git importYour existing project, or a new one
Solana referencesSolana-specific training data and templatesOffline reference for 50 programs and 1,559 instructions, a pinned Solana documentation set, error decoding and toolchain checks
Test chainA solana-test-validator in each workspaceA local Surfpool chain, with /chain, /airdrop, /account and /rewind
When work is doneYou review each AI change and apply itThe model can only propose; a separate check runs your requirements on a fresh offline chain
DeployOne-click deploy to localhost, devnet, testnet or mainnetWorks against the local chain; mainnet work comes back to you as a transaction to review and sign
KeysAn auto-generated wallet in each workspaceOn macOS, a Seatbelt sandbox blocks commands from reading ~/.config/solana and ~/.ssh

How does each one write a Solana program?

Código, from its documentation:

  1. You describe the program in the workspace, or start from a template.
  2. The agent generates the program, tests and frontend, and can update several files at once.
  3. You review the changes, then simulate instructions, build and deploy from the browser.

Oh My Sol, from its docs:

  1. You describe the work. OMS creates a task with a requirements file and its own Git worktree.
  2. The model reads and edits files in that worktree and runs commands in the sandbox, with Solana references and error decoding available as tools.
  3. When it thinks it is finished, its only option is to propose that it is done.
  4. A separate check runs your requirements on a fresh offline Surfpool chain. A pass closes the task; a fail sends it back for another attempt while budget remains.

With the same model, OMS beats both Codex CLI and Claude Code on Solana Bench and fresh Solana tasks. The benchmarks overview has the numbers.

Which should I use?

  • Código if you want a browser IDE with no local setup, templates for common programs and one-click deploy.
  • Oh My Sol if you work in a terminal on your own project, want to keep the model you already pay for, and want the work checked on a local chain before it counts as done.

Frequently asked questions

What is the main difference between Oh My Sol and Código?

Código is a browser IDE with its own Solana-trained model, templates and one-click deploy. Oh My Sol (OMS) is a terminal harness that runs on your machine with the model you already use, keeps Solana references, error decoding and toolchain checks next to it, and closes a task only when your requirements pass on a fresh local Surfpool chain.

Which one should a Solana beginner use?

Código is designed to need no local setup: Anchor, the Solana CLI and Cargo come preinstalled in each workspace, and templates cover common programs. OMS needs Node 22.19.0 or newer and Git, and its first command downloads Surfpool and checks your toolchain. It suits developers who want to work on their own project in a terminal.

Which models do they use?

Código says it is powered by its own custom-trained model built for Solana. OMS does not supply a model; you connect OpenAI, Anthropic, Xiaomi MiMo or OpenCode access, including a ChatGPT or Claude subscription.

Can either one deploy to mainnet?

Código offers one-click deploy to localhost, devnet, testnet or mainnet from the browser. OMS never holds your mainnet keys; it works against a local Surfpool chain, and mainnet work comes back to you as a transaction to review and sign yourself.

Try it on your own program

Oh My Sol 0.1.0 runs on macOS and Linux, and on Windows through WSL 2. You need Node 22.19.0 or newer and Git. You bring your own model access: an API key, a ChatGPT subscription or a Claude subscription.

npx @barestate/oms doctor --install
npx @barestate/oms