Thinking Machines

透過 Mastra 的模型路由器存取 2 個 Thinking Machines 模型。系統會自動使用 TINKER_API_KEY 環境變數完成身分驗證。

如需進一步了解，請參閱Thinking Machines 文件。

.env

TINKER_API_KEY=your-api-key

src/mastra/agents/my-agent.ts

import { Agent } from '@mastra/core/agent'

const agent = new Agent({
  id: 'my-agent',
  name: 'My Agent',
  instructions: 'You are a helpful assistant',
  model: 'thinkingmachines/thinkingmachines/Inkling',
})

// Generate a response
const response = await agent.generate('Hello!')

// Stream a response
const stream = await agent.stream('Tell me a story')
for await (const chunk of stream) {
  console.log(chunk)
}

資訊

Mastra 使用與 OpenAI 相容的 /chat/completions endpoint。部分 Provider 特有功能可能無法使用，詳情請參閱Thinking Machines 文件。

模型

2 available models

Model	Context	Tools	Reasoning	Image	Audio	Video	Input $/1M	Output $/1M
thinkingmachines/thinkingmachines/Inkling	66K						$2	$5
thinkingmachines/thinkingmachines/Inkling:peft:262144	262K						$4	$9

進階設定

自訂標頭

src/mastra/agents/my-agent.ts

const agent = new Agent({
  id: 'custom-agent',
  name: 'custom-agent',
  model: {
    url: 'https://tinker.thinkingmachines.dev/services/tinker-prod/anthropic/api/v1',
    id: 'thinkingmachines/thinkingmachines/Inkling',
    apiKey: process.env.TINKER_API_KEY,
    headers: {
      'X-Custom-Header': 'value',
    },
  },
})

動態選擇模型

src/mastra/agents/my-agent.ts

const agent = new Agent({
  id: 'dynamic-agent',
  name: 'Dynamic Agent',
  model: ({ requestContext }) => {
    const useAdvanced = requestContext.task === 'complex'
    return useAdvanced
      ? 'thinkingmachines/thinkingmachines/Inkling:peft:262144'
      : 'thinkingmachines/thinkingmachines/Inkling'
  },
})

直接安裝 Provider

你也可以直接將此 Provider 安裝為獨立套件，用它取代 Mastra 模型路由字串。詳情請參閱套件文件。

npm

npm install @ai-sdk/anthropic

pnpm

pnpm add @ai-sdk/anthropic

Yarn

yarn add @ai-sdk/anthropic

Bun

bun add @ai-sdk/anthropic