Getting Started

Configure AI

Connect an AI model to unlock deeper, context-aware analysis and the one-approval fix workflow. DeployReady supports Claude, OpenAI, and local Ollama models.

What AI adds

  • Deeper, context-aware explanations of each finding
  • AI-proposed fixes with before/after diffs
  • Architecture and best-practice suggestions

The core scan runs with or without AI

The 30+ structured checks run without a model using --no-ai. Connecting a model adds AI explanations and the approve-a-diff fix workflow on top.

Option 1: Claude (Anthropic)

Use Anthropic's Claude models for the richest analysis. Run the init command and select Claude, then paste your API key when prompted.

terminal
npx deployready@latest init

> Select model provider:
   1  Claude (Anthropic)
   2  OpenAI
   3  Ollama (local)

> 1
> Paste your Anthropic API key: ************

Your key stays local

The API key is stored locally with restricted permissions (0600) at ~/.deployready/config.json. Only redacted, structured findings are ever sent to the model — never your raw source code.

Option 2: OpenAI

Prefer OpenAI? Run the init command, select OpenAI, and paste your API key. GPT-4o and newer models are supported.

terminal
npx deployready@latest init

> Select model provider:
   1  Claude (Anthropic)
   2  OpenAI
   3  Ollama (local)

> 2
> Paste your OpenAI API key: ************

Option 3: Ollama (fully local)

Prefer to keep everything on your machine? Use Ollama to run a local model. Nothing leaves your network.

terminal
# Pull a model with Ollama first
ollama pull llama3

# Then configure DeployReady
npx deployready@latest init
> Select: Ollama
> Model: llama3

Privacy model

What gets sent to AI

  • Structured findings as JSON (severity, type, file, line)
  • All secret values redacted before sending
  • Project metadata (file count, routes, detected stack)

What never leaves your machine

  • Raw source code
  • Full secret or credential values
  • Environment variable values

Switching providers or changing your key

Already set up and want to change provider, swap in a new API key, or pick a different model? You have a few options:

  • Re-run deployready init and choose again.
  • From inside a session, open the menu and pick Set up or change the AI model.
  • Set ANTHROPIC_API_KEY or OPENAI_API_KEY — env vars override the config file and are never written to disk.

Full configuration reference

For every config field, default values, and the config command, see the CLI Reference.

Disable AI per run

Even when configured, you can skip AI on any individual run.

terminal
npx deployready@latest analyze . --no-ai

AI suggestions are not guarantees

Models can be wrong. Always review AI-proposed fixes and explanations against your own judgment. See the Security Disclaimer.
PreviousDemoNextStatic Analysis