Guide d’intégration

Vercel AI SDK + OrcaRouter

The Vercel AI SDK is the de-facto way to build streaming LLM apps in Next.js, SvelteKit, and Nuxt. Its OpenAI provider accepts a custom baseURL — wire it to OrcaRouter and ship every model on the catalog with one key.

Obtenir une clé API Vercel AI SDK site web ↗

Étapes de configuration

Brancher OrcaRouter en 5 minutes

1.Install: npm install ai @ai-sdk/openai
2.Set ORCAROUTER_API_KEY in your .env.local
3.Use createOpenAI({ baseURL, apiKey }) with baseURL=https://api.orcarouter.ai/v1
4.Call streamText / generateText / generateObject as usual.
5.Every streaming token routes through the cheapest healthy backend.

Exemple de configuration

import { createOpenAI } from '@ai-sdk/openai';
import { streamText } from 'ai';

const orca = createOpenAI({
  baseURL: 'https://api.orcarouter.ai/v1',
  apiKey: process.env.ORCAROUTER_API_KEY,
});

const result = await streamText({
  model: orca('claude-sonnet-4'),
  prompt: 'Write a haiku about TypeScript.',
});

// In a Next.js route handler:
return result.toDataStreamResponse();

Pourquoi router Vercel AI SDK via OrcaRouter ?

The Vercel AI SDK's model abstraction already gives you portability across OpenAI, Anthropic, etc. Adding OrcaRouter on top means you don't need to manage N API keys or hit N different rate limits — one key, one bill, automatic failover under the hood.

Autres intégrations

⌘ Cursor

AI-first code editor.

🤖 Cline

Autonomous coding agent for VS Code.

➡ Continue

Open-source AI code assistant for VS Code and JetBrains.

✎ Aider

AI pair programming in your terminal.

🦜 LangChain

Compose LLM apps in Python or JS.

🦙 LlamaIndex