PAWS SDK

A drop-in replacement for the Anthropic SDK that works with OpenRouter, Ollama, vLLM, OpenAI, and any OpenAI-compatible API.

Installation

npm install @paws/paws-sdk

Quick Start

import { Anthropic } from '@paws/paws-sdk'

const client = new Anthropic({
  provider: 'openrouter',  // or 'ollama', 'vllm', 'openai', 'custom'
  apiKey: process.env.OPENROUTER_API_KEY,
  model: 'anthropic/claude-sonnet-4',
})

const response = await client.messages.create({
  model: 'anthropic/claude-sonnet-4',
  max_tokens: 1024,
  messages: [{ role: 'user', content: 'Hello!' }],
})

Configuration

| Environment Variable | Description | |---|---| | PAWS_PROVIDER | Provider type: openrouter, ollama, vllm, openai, custom (default: openrouter) | | OPENROUTER_API_KEY | API key for OpenRouter | | OPENAI_API_KEY | API key for OpenAI (fallback provider) | | PAWS_MODEL | Default model to use | | OPENROUTER_BASE_URL | Custom OpenRouter base URL | | OPENAI_BASE_URL | Custom OpenAI base URL |

API

createMessage(params)

Non-streaming message completion.

import { createMessage } from '@paws/paws-sdk'

const msg = await createMessage({
  model: 'anthropic/claude-sonnet-4',
  max_tokens: 1024,
  messages: [{ role: 'user', content: 'Hello!' }],
  tools: [{ name: 'search', description: '...', input_schema: {} }],
})

streamMessage(params)

Streaming message completion with server-sent events.

import { streamMessage } from '@paws/paws-sdk'

for await (const event of streamMessage({
  model: 'anthropic/claude-sonnet-4',
  max_tokens: 1024,
  messages: [{ role: 'user', content: 'Tell me a story' }],
})) {
  if (event.type === 'content_block_delta' && event.delta?.text) {
    process.stdout.write(event.delta.text)
  }
}

configurePaws(opts)

Configure global provider options.

import { configurePaws } from '@paws/paws-sdk'

configurePaws({
  provider: 'ollama',
  baseURL: 'http://localhost:11434/v1',
  model: 'llama3',
})

Providers

| Provider | baseURL | apiKey | |---|---|---| | openrouter | https://openrouter.ai/api/v1 | OPENROUTER_API_KEY | | ollama | http://localhost:11434/v1 | (optional) | | vllm | Your vLLM endpoint | (optional) | | openai | https://api.openai.com/v1 | OPENAI_API_KEY | | custom | Set via baseURL | Set via apiKey |

Build

# Install dependencies
npm install

# Build the CLI bundle
npm run build

# Production build (minified)
npm run build:prod

# Watch mode
npm run build:watch

Publishing

npm publish --access public