DeepSeek

Accédez à 4 modèles DeepSeek via le routeur de modèles de Mastra. L’authentification est gérée automatiquement à l’aide de la variable d’environnement DEEPSEEK_API_KEY.

Pour en savoir plus, consultez la documentation DeepSeek.

.env

DEEPSEEK_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: "deepseek/deepseek-chat"
});

// 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);
}

Modèles

4 available models

Model	Context	Tools	Reasoning	Image	Audio	Video	Input $/1M	Output $/1M
deepseek/deepseek-chat	1.0M						$0.14	$0.28
deepseek/deepseek-reasoner	1.0M						$0.14	$0.28
deepseek/deepseek-v4-flash	1.0M						$0.14	$0.28
deepseek/deepseek-v4-pro	1.0M						$0.43	$0.87

Configuration avancée

En-têtes personnalisés

src/mastra/agents/my-agent.ts

const agent = new Agent({
  id: "custom-agent",
  name: "custom-agent",
  model: {
    url: "https://api.deepseek.com",
    id: "deepseek/deepseek-chat",
    apiKey: process.env.DEEPSEEK_API_KEY,
    headers: {
      "X-Custom-Header": "value"
    }
  }
});

Sélection dynamique du modèle

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
      ? "deepseek/deepseek-v4-pro"
      : "deepseek/deepseek-chat";
  }
});

Options du Provider

DeepSeek prend en charge les options propres au Provider suivantes via le paramètre providerOptions :

const response = await agent.generate("Hello!", {
  providerOptions: {
    deepseek: {
      // See available options in the table below
    }
  }
});

Options disponibles

thinking?:

{ type?: "adaptive" | "enabled" | "disabled" | undefined; } | undefined

reasoningEffort?:

"low" | "medium" | "high" | "xhigh" | "max" | undefined

strictJsonSchema?:

boolean | undefined