LLM API for TypeScript

Use the TypeScript client when you want a minimal wrapper around the same API your product uses in production. It keeps your app code focused on workflow logic instead of provider-specific plumbing.

npm install llmwise

Initialize the client with your `mm_sk_...` key and reuse it anywhere you need chat, compare, or routed requests. One key gives you access to the full multi-model surface.

import { LLMWise } from "llmwise";

const client = new LLMWise("mm_sk_...");

Keep TypeScript integrations simple by defaulting to `model: "auto"`. You can still pin a specific model later, but Auto is the quickest way to get working behavior with sensible cost control.

const response = await client.chat({
  model: "auto",
  messages: [{ role: "user", content: "Classify this support ticket." }],
});

When you are deciding between providers for a feature, compare mode helps you test model quality without switching SDKs or auth flows. This is especially useful in TypeScript products where experimentation speed matters.

const comparison = await client.compare({
  model: "auto",
  messages: [{ role: "user", content: "Write release notes for this patch." }],
});

As your app grows, keep the same TypeScript client while changing routing and billing policy in LLMWise. That is the main advantage of a unified API layer: your application code stays stable while your infrastructure choices evolve.