MegaNova JavaScript SDK

Official JavaScript/TypeScript SDK for the MegaNova AI platform — inference, cloud agents, and more.

Installation

npm install meganova

Requires Node.js 18+ (uses native fetch).

Quick Start

Chat Completion

import { MegaNova } from "meganova";

const client = new MegaNova({ apiKey: "YOUR_API_KEY" });

const response = await client.chat.completions.create({
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "What is the capital of France?" },
  ],
  model: "meganova-ai/manta-flash-1.0",
});

console.log(response.choices[0].message.content);

Streaming

const stream = await client.chat.completions.create({
  messages: [{ role: "user", content: "Write a haiku about space." }],
  model: "meganova-ai/manta-flash-1.0",
  stream: true,
});

for await (const chunk of stream) {
  const content = chunk.choices[0]?.delta?.content;
  if (content) process.stdout.write(content);
}

Cloud Agents

Chat with deployed MegaNova Studio agents using just an API key — no authentication setup required.

import { CloudAgent } from "meganova";

const agent = new CloudAgent({ apiKey: "agent_xxx..." });

// Get agent info
const info = await agent.info();
console.log(`${info.name}: ${info.welcome_message}`);

// Chat
const response = await agent.chat("Hello!");
console.log(response.response);

Multi-turn Conversation

const conv = agent.conversation();
const r1 = await conv.chat("I need help with my account");
const r2 = await conv.chat("My email is [email protected]");
// conversation_id is tracked automatically

Tool Confirmation

Some agents use tools (ticket creation, email, etc.) that require approval:

const conv = agent.conversation();
const response = await conv.chat("Create a support ticket for my login issue");

if (conv.pendingToolCall) {
  console.log(`Agent wants to: ${conv.pendingToolCall.description}`);
  const result = await conv.confirm(); // or conv.reject()
}

OpenAI-compatible Completions

// Non-streaming
const response = await agent.completions({
  messages: [{ role: "user", content: "Hello!" }],
});
console.log(response.choices[0].message.content);

// Streaming
const stream = await agent.completions({
  messages: [{ role: "user", content: "Tell me a story" }],
  stream: true,
});

for await (const chunk of stream) {
  const content = chunk.choices[0]?.delta?.content;
  if (content) process.stdout.write(content);
}

Image Generation

const response = await client.images.generate({
  prompt: "A futuristic city skyline at sunset, digital art",
  model: "ByteDance/SeedDream-4-5",
  width: 1024,
  height: 1024,
});

// response.data[0].b64_json contains the base64-encoded image

Model Discovery

// List all models
const models = await client.models.list();

// List serverless models with pricing
const textModels = await client.serverless.listModels("text_generation");
for (const model of textModels.models) {
  console.log(`${model.model_name}: $${model.cost_per_1k_input}/1K input`);
}

// Image models
const imageModels = await client.serverless.listModels("text_to_image");

Supported Models

• Text: meganova-ai/manta-mini-1.0, manta-flash-1.0, manta-pro-1.0, Qwen/Qwen3-235B-A22B-Instruct-2507
• Image: ByteDance/SeedDream-4-5, black-forest-labs/FLUX.1-dev
• Audio: Systran/faster-whisper-large-v3
• Vision: Qwen/Qwen2.5-VL-7B-Instruct

Use client.serverless.listModels() for the full list with pricing.

Error Handling

import { MeganovaError, AuthenticationError, RateLimitError } from "meganova";

try {
  await client.chat.completions.create({ ... });
} catch (err) {
  if (err instanceof AuthenticationError) {
    console.error("Invalid API key");
  } else if (err instanceof RateLimitError) {
    console.error("Rate limited, retry later");
  } else if (err instanceof MeganovaError) {
    console.error(`API error (${err.status}): ${err.message}`);
  }
}

Configuration

const client = new MegaNova({
  apiKey: "YOUR_API_KEY",
  baseUrl: "https://api.meganova.ai/v1", // default
  timeout: 60000,  // ms, default 60s
  maxRetries: 2,   // default
});

Environment variable: set MEGANOVA_API_KEY and read it in your app.

Examples

See the examples/ directory for runnable scripts.

License

MIT