一个接口，
接入所有主流大模型

$ tai chat --model gpt-4 \
    --system "You are a poet." \
    --message "写首关于终端的诗" --stream

光标一闪，
问题落进黑色的屏幕；
回车之后，
万千模型同时应答。

$ 

$ tai image generations --model dall-e-3 \
    --prompt "A cat in a spacesuit" \
    --size 1024x1024 --out cat.png
Saved to cat.png

$ 

$ tai video generations --model kling-video-o1 \
    --prompt "A sunset over the sea" \
    --first-frame ./start.png --audio \
    --out sunset.mp4
Saved to sunset.mp4

$ 

$ tai audio speech --model tts-1 \
    --input "你好，欢迎使用 ThinkAI" \
    --voice alloy --out hello.mp3
Saved to hello.mp3

$ 

$ tai music bgm --model [MODEL] \
    --prompt "史诗管弦" --duration 60 \
    --out bgm.mp3
Saved to bgm.mp3

$ 

$ tai model list --type chat
[
  { type: 'chat', code: 'gpt-4', … },
  …
]

$ 

use think\ai\Client;

$client = new Client('YOUR_TOKEN');

$result = $client->chat()->completions([
    'model'    => 'gpt-4',
    'messages' => [['role' => 'user', 'content' => 'Hello!']],
    'stream'   => true,
]);

foreach ($result as $chunk) {
    dump($chunk);
}

$result = $client->decision()->evaluate([
    'model'     => 'jev-latest',
    'state'     => 'Help! My payouts have been failing for 3 days.',
    'questions' => [
        'is_urgent'  => ['type' => 'noul', 'instructions' => 'Does this convey urgency?'],
        'department' => ['type' => 'choice', 'criteria' => ['billing' => '…', 'technical' => '…']],
    ],
]);

// $result['answers']['is_urgent']['noul']     => 0.95
// $result['answers']['department']['choice']  => 'billing'

import { Client } from '@topthink/ai';

const client = new Client({ apiKey: 'sk-...' });

const stream = await client.chat.completions({
  model: 'gpt-4',
  messages: [{ role: 'user', content: 'Hello!' }],
  stream: true,
});

for await (const event of stream) {
  process.stdout.write(event.delta.content);
}

const result = await client.image.generations({
  model: 'dall-e-3',
  prompt: 'A cat in a spacesuit',
  size: '1024x1024',
});

console.log(result.images[0].url);