Vercel AI SDK

Meridian with the Vercel AI SDK

Drop Meridian into any Vercel AI SDK project. One baseURL change, zero vendor lock-in. Use the same streaming, tool-calling, and structured output APIs you already know.

Get an API keyRead the docs

Quick start

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

const meridian = createOpenAI({
  baseURL: 'https://api.getnimbus.net/v1',
  apiKey: process.env.MERIDIAN_API_KEY,
});

const { text } = await generateText({
  model: meridian('meridian-pro'),
  prompt: 'Explain zero-knowledge proofs',
});

What you get

Streaming

streamText works out of the box. Real-time token delivery with zero config.

Tool calling

Define tools with zod schemas. Meridian calls them and returns structured JSON.

Structured output

generateObject with zod gives you typed, validated responses every time.

Edge runtime

Deploy on Vercel Edge. Meridian responds fast from global inference nodes.

No vendor lock

Swap between Meridian, OpenAI, and Anthropic by changing one import.

TypeScript first

Full type safety from prompt to parsed output. Autocomplete in your IDE.

Ready to ship?

Install the SDK, set your baseURL, and start building. Same API, better inference.

Start freeView full docs