Log inCreate account

Docs

Getting started

On your machine

  1. Install the toolchain. Install Node.js 22, pnpm 11.9.0, Python 3.11, and Git.
  2. Install Ollama. Install Ollama and run ollama serve, then ollama pull qwen2.5-coder:7b. qwen2.5-coder:1.5b is the smaller pull when RAM is tight.
  3. Build the repo. From the repo: pnpm install, pnpm build, then .\start.ps1 on Windows.
  4. Run the CLI. pnpm codesync scan ., diagnose, repair . --apply, review, and generate "Create a React status component" --apply.

On codesyncsi.com

  1. Create an account. Create an account and verify the email. In local development the message is stored in the outbox.
  2. Subscribe. Subscribe to Standard, Expert, or Extreme. The server rejects projects and runs without an active subscription.
  3. Add a project. Upload a zip, or on Expert and Extreme connect owner/name with a GitHub token.
  4. Run the engine. Run scan, diagnose, repair, review, or generate. Download the zip from the run.

Model ids default to Llama 3.2 3B, Llama 3.1 8B, and Llama 3.3 70B. If Workers AI does not list one of those ids, set the HOSTED_MODEL_* variable to an id your account can run. The run stores the error instead of inventing a completion.

Next: Ollama locally, pricing, and the glossary.