metrix.ai Node.js SDK

LLM Cost Intelligence SDK — track, optimize, and control your AI spend across OpenAI, Anthropic, and Gemini.

Install

npm install metrix-ai

Quickstart

import { MetrixClient } from "metrix-ai";

const client = new MetrixClient({
  apiKey: "your-metrix-token",
  providerApiKey: "your-anthropic-key",
  project: "my-project",
  debug: true,
});

const response = await client.chat({
  provider: "anthropic",
  model: "claude-sonnet-4-6",
  messages: [{ role: "user", content: "Hello from metrix.ai!" }],
});

console.log(response.content);
console.log(`Tokens used: ${response.totalTokens}`);
console.log(`Latency: ${response.latencyMs}ms`);

Environment Variables

export METRIX_API_KEY=your-metrix-token
export METRIX_PROVIDER_API_KEY=your-anthropic-key
export METRIX_PROJECT=my-project
export METRIX_AUTO_ROUTE=true
export METRIX_DEBUG=true

Then just:

import { MetrixClient } from "metrix-ai";
const client = new MetrixClient();

Smart Routing

const client = new MetrixClient({
  apiKey: "...",
  providerApiKey: "...",
  autoRoute: true,
});

const response = await client.chat({
  provider: "anthropic",
  model: "claude-opus-4-6",
  messages: [{ role: "user", content: "What is 2+2?" }],
});

if (response.routed) {
  console.log(`Routed from ${response.originalModel} → ${response.model}`);
}

Project Attribution

const response = await client.chat({
  provider: "anthropic",
  model: "claude-sonnet-4-6",
  messages: [{ role: "user", content: "Summarize this..." }],
  project: "customer-support",
  tags: { env: "prod", feature: "summarization" },
});

Supported Providers

| Provider | Models | |-----------|--------| | anthropic | claude-opus-4-6, claude-sonnet-4-6, claude-haiku-4-5 | | openai | gpt-4o, gpt-4o-mini, gpt-4-turbo | | gemini | gemini-1.5-pro, gemini-1.5-flash |

Error Handling

import {
  MetrixClient,
  MetrixAuthError,
  ProviderRateLimitError,
  ProviderTimeoutError,
  MetrixConnectionError,
  UnsupportedProviderError,
} from "metrix-ai";

try {
  const response = await client.chat({ ... });
} catch (err) {
  if (err instanceof MetrixAuthError) console.log("Invalid token");
  if (err instanceof ProviderRateLimitError) console.log("Rate limited");
  if (err instanceof ProviderTimeoutError) console.log("Timed out");
  if (err instanceof MetrixConnectionError) console.log("Proxy unreachable");
  if (err instanceof UnsupportedProviderError) console.log(err.message);
}

Configuration

| Option | Type | Default | Description | |--------|------|---------|-------------| | apiKey | string | env | metrix.ai JWT token | | providerApiKey | string | env | LLM provider API key | | project | string | undefined | Project slug | | autoRoute | boolean | false | Enable smart routing | | tags | object | undefined | Attribution tags | | timeout | number | 60000 | Timeout in ms | | maxRetries | number | 2 | Retry attempts | | debug | boolean | false | Colored debug output | | baseUrl | string | localhost | Proxy base URL |

License

MIT — Ahmad Sharique