TokenLab AI SDK Provider

TokenLab provider for Vercel AI SDK, built on the official OpenAI-compatible provider package.

TokenLab supports OpenAI-compatible /v1 routes plus native Responses, Anthropic Messages, Gemini, media, audio, embeddings, rerank, and translation APIs. This package focuses on the AI SDK OpenAI-compatible path for text generation, streaming, tools, and structured output.

Install

npm install ai @tokenlabai/ai-sdk-provider

Version 0.2.x targets AI SDK 7 and Node.js 22 or newer.

Usage

import { generateText } from "ai";
import { tokenlab } from "@tokenlabai/ai-sdk-provider";

const { text } = await generateText({
  model: tokenlab.chatModel("gpt-5.5"),
  prompt: "Explain TokenLab in one sentence."
});

console.log(text);

Custom Provider Instance

import { createTokenLab } from "@tokenlabai/ai-sdk-provider";

export const tokenlab = createTokenLab({
  apiKey: process.env.TOKENLAB_API_KEY,
  baseURL: "https://api.tokenlab.sh/v1"
});

Environment

TOKENLAB_API_KEY=sk-your-tokenlab-key

Model Discovery

Prefer live model discovery for production routing:

curl https://api.tokenlab.sh/v1/models
curl "https://api.tokenlab.sh/v1/models?recommended_for=embedding"
curl https://api.tokenlab.sh/llms.txt

Native Endpoints

Use this provider for AI SDK OpenAI-compatible text paths. For TokenLab native endpoint families, use TokenLab's REST API directly:

• Responses: POST https://api.tokenlab.sh/v1/responses
• Anthropic Messages: POST https://api.tokenlab.sh/v1/messages
• Gemini generateContent: POST https://api.tokenlab.sh/v1beta/models/{model}:generateContent
• Images, videos, audio, embeddings, rerank, and translation: see https://docs.tokenlab.sh

Links

• Docs: https://docs.tokenlab.sh/integrations/vercel-ai-sdk
• OpenAPI: https://docs.tokenlab.sh/openapi.json
• Model catalog: https://api.tokenlab.sh/v1/models
• MCP server: https://github.com/hedging8563/tokenlab-mcp-server