deepseek-v4

n8n community node for DeepSeek chat models (Pro / Flash) with reasoning_content (Chain-of-Thought) support, usable as a LangChain sub-node inside the AI Agent and Basic LLM Chain.

Documentation

n8n-nodes-deepseek

Community n8n node that exposes DeepSeek chat models — including DeepSeek V4 Pro and DeepSeek V4 Flash — as a LangChain chat-model sub-node, so you can drop it into the AI Agent or Basic LLM Chain the same way as the OpenAI / OpenRouter chat-model nodes.

DeepSeek's thinking mode is supported out of the box: the chain-of-thought is captured into reasoning_content and surfaced on the AI message as additional_kwargs.reasoning_content, ready to be used by downstream nodes.

Features

  • Models: deepseek-v4-pro, deepseek-v4-flash, deepseek-reasoner, deepseek-chat, plus a free-form Custom option for future models.
  • Thinking-mode toggle (on by default) and reasoning_effort control (high / max).
  • Streaming and non-streaming both surface reasoning_content.
  • Tool calling supported (DeepSeek V4 thinking mode supports tools).
  • Drop-in replacement for lmChatOpenRouter / lmChatOpenAi in any AI Agent workflow.

Install

In your n8n instance:

  1. Settings → Community NodesInstall.
  2. Enter n8n-nodes-deepseek.
  3. Reload, then add credentials of type DeepSeek API (your key from https://platform.deepseek.com/api_keys).
  4. Add the DeepSeek Chat Model node and wire its Model output into an AI Agent node.

Reading the chain-of-thought

After the AI Agent runs, the assistant message has both:

  • output (or content) — the final answer.
  • additional_kwargs.reasoning_content — the model's chain-of-thought.

Example workflow JSON snippet:

{
  "nodes": [
    {
      "parameters": {
        "model": "deepseek-v4-pro",
        "options": { "thinking": true, "reasoningEffort": "high" }
      },
      "type": "n8n-nodes-deepseek.lmChatDeepSeek",
      "typeVersion": 1,
      "position": [1312, 1824],
      "id": "deepseek-pro",
      "name": "DeepSeek Chat Model",
      "credentials": {
        "deepSeekApi": { "id": "REPLACE", "name": "My DeepSeek key" }
      }
    }
  ],
  "connections": {
    "DeepSeek Chat Model": { "ai_languageModel": [[]] }
  }
}

Develop / publish

cd packages/n8n-nodes-deepseek
npm install
npm run build
# verify the dist/ folder, then:
npm publish --access public

Make sure package.json has your real name, author, and repository URL before publishing.

License

MIT

Discussion