Vercel AI SDK with 6bee.ai

Before you start

• A Node.js project.
• Create an account with a verified email address, add prepaid balance, then create a key on the console's API keys page. Keep the key secret.
• Model IDs are listed in the console and returned by GET https://api.6bee.ai/v1/models. Replace <model id> with one of them.

Install and create the provider

npm install ai @ai-sdk/openai-compatible

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

const sixbee = createOpenAICompatible({
  name: 'sixbee',
  baseURL: 'https://api.6bee.ai/v1',
  apiKey: process.env.SIXBEE_API_KEY,
});

const result = streamText({
  model: sixbee('<model id>'),
  prompt: 'Explain this stack trace',
});
for await (const text of result.textStream) process.stdout.write(text);

Good to know

• Call the API from server code only: a route handler, a server action or a script. A key in browser code is public.
• An account runs 2 requests at a time. If your app serves several users from one key, queue the calls or expect 429.
• On 429, wait for the number of seconds in the Retry-After header.

Context

Input plus output up to the model’s window: 128K to 1M tokens, depending on the model.

Output

max_tokens up to the model’s output cap: 16,384 to 131,072, depending on the model. Larger values are lowered to the cap, not rejected. If omitted: 8,192 to 32,768.

Input tokens

1,000,000 uncached tokens per rolling minute. Cache hits don’t count.

Requests

60 per minute.

Concurrency

2 requests at a time.

Request body

Up to 8 MiB. n must be 1.

Troubleshooting

HTTP 401

The key is missing or wrong. Copy it again from the console, without spaces or quotes.

Model not found

The model ID does not match. List the valid IDs with GET /v1/models or check the console.

HTTP 429

You are over a per-account limit (requests, input tokens or concurrency). The Retry-After header says how long to wait.

HTTP 400 on long sessions

The request no longer fits the model’s context window. Compact or clear the conversation, or switch to a model with a larger window.

Request rejected for balance

Each request first reserves its input plus max_tokens of output. Top up, or lower max_tokens.