Configuration

Config File

Create nudge.config.json in your project root, or run npx @nudge-ai/cli init to generate one interactively:

nudge.config.json

{
  "ai": {
    "provider": "openrouter",
    "apiKeyEnvVar": "OPENROUTER_API_KEY",
    "model": "anthropic/claude-sonnet-4"
  }
}

Option	Default	Description
`generatedFile`	`src/prompts.gen.ts`	Output path for generated file
`promptFilenamePattern`	`*/.prompt.{ts,js}`	Glob pattern for prompt files
`ai.provider`	—	`"openai"`, `"openrouter"`, or `"local"`
`ai.apiKeyEnvVar`	—	Environment variable name (optional for `"local"`)
`ai.model`	—	Model identifier
`ai.baseUrl`	—	Custom API base URL (required for `"local"`)

OpenRouter — openrouter.ai

{
  "ai": {
    "provider": "openrouter",
    "apiKeyEnvVar": "OPENROUTER_API_KEY",
    "model": "anthropic/claude-sonnet-4"
  }
}

OpenAI — platform.openai.com

{
  "ai": {
    "provider": "openai",
    "apiKeyEnvVar": "OPENAI_API_KEY",
    "model": "gpt-4o"
  }
}

Local — For local models (llama.cpp, Ollama, etc.) with OpenAI-compatible API:

{
  "ai": {
    "provider": "local",
    "baseUrl": "http://localhost:11434/v1",
    "model": "llama2"
  }
}

npx @nudge-ai/cli generate --force  # Force regenerate all prompts