meshs-one · Node.js SDK

One API key. 30+ AI models. Zero dependencies.

Why

• 🎯 30+ models, one key — GPT-5, Claude 4, Gemini 2.5, DeepSeek R2, Qwen 3
• 💰 Up to 80% cheaper than official pricing
• 🔄 OpenAI SDK compatible — swap base_url, zero migration
• 🛡️ Built-in retry — auto-retries on network errors
• 🧵 Streaming — async generator for real-time responses
• 📦 Zero dependencies — pure Node.js, no npm bloat

Install

npm install meshs-one

Quick Start

const MeshsOne = require('meshs-one');

const client = new MeshsOne({ apiKey: 'your-api-key' });

// Simple chat
const reply = await client.chat('Explain quantum computing in one sentence.');
console.log(reply);

// With system prompt + model selection
const reply2 = await client.chat([
  { role: 'system', content: 'You are a Python expert.' },
  { role: 'user', content: 'Write a FastAPI endpoint for file upload.' }
], { model: 'claude-4-sonnet' });

console.log(reply2);

API Reference

new MeshsOne(options)

Option	Type	Default	Description
apiKey	string	required	Your API key from api.meshs.one
timeout	number	60000	Request timeout (ms)
maxRetries	number	3	Auto-retry count

client.chat(messages, options?)

Send a chat completion. Returns the response text (string).

// String shorthand
await client.chat('Hello!');

// Full message array
await client.chat([
  { role: 'user', content: 'Hello!' }
]);

// With options
await client.chat('Hello!', {
  model: 'claude-4-opus',
  temperature: 0.7,
  maxTokens: 2000
});

// Streaming
const stream = await client.chat('Tell me a story', { stream: true });
for await (const chunk of stream) {
  process.stdout.write(chunk);
}

client.listModels()

const models = await client.listModels();
// [{ id: 'gpt-5', owned_by: 'openai' }, ...]

client.ping()

const ok = await client.ping(); // true | false

client.getKnownModels()

const models = client.getKnownModels();
// ['claude-4-opus', 'claude-4-sonnet', 'gpt-5', 'gpt-4o', ...]

Error Handling

const { MeshsError, MeshsAuthError, MeshsRateLimitError } = MeshsOne;

try {
  await client.chat('Hello');
} catch (err) {
  if (err instanceof MeshsAuthError) {
    console.error('Invalid API key');
  } else if (err instanceof MeshsRateLimitError) {
    console.error('Too many requests, slow down');
  } else if (err instanceof MeshsError) {
    console.error(`Meshs error: ${err.message} (HTTP ${err.status})`);
  }
}

OpenAI SDK Compatibility

Already using openai npm package? Just change baseURL:

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://api.meshs.one/v1',
  apiKey: 'your-api-key'
});
// All your existing code works!

Models

Model	Provider	Best for
gpt-5	OpenAI	General purpose, math
gpt-4o	OpenAI	Balanced cost/quality
claude-4-opus	Anthropic	Long-form writing
claude-4-sonnet	Anthropic	Code generation
gemini-2.5-pro	Google	Translation, vision
gemini-2.5-flash	Google	Fast, cheap
deepseek-r2	DeepSeek	Math, reasoning
deepseek-v3	DeepSeek	Budget batch
qwen-3	Alibaba	Budget batch
qwen-3-turbo	Alibaba	Fast, cheapest

📚 Tutorials

Date	Tutorial	Blog	Code
2026-06-24	AI API Gateway Quickstart: One Key, 30+ Models in 5 Minutes	blog.meshs.one	examples/quickstart.js · smart-router.js

More tutorials coming soon. Have a use case to share? Open an issue.

License

MIT © Meshs One

---

Get API Key · Blog · X