---
title: "DeepSeek Integration"
source: https://docs.autohand.ai/integrations/deepseek
---

# DeepSeek

Use DeepSeek V4 Flash and V4 Pro directly from Autohand through DeepSeek's OpenAI-compatible API.

## Overview

DeepSeek is a good fit for coding, reasoning, and cost-sensitive agent sessions. Autohand supports the `deepseek` provider with API-key authentication and model switching through `/model`.

## Setup

1.  Create a DeepSeek API key from the DeepSeek platform.
2.  Set `DEEPSEEK_API_KEY` or save the key in Autohand config.
3.  Select a DeepSeek model during setup or with `/model`.

``` bash
export DEEPSEEK_API_KEY="your-deepseek-api-key"

autohand --provider deepseek --model deepseek-v4-flash
```

## CLI configuration

``` json
{
  "provider": "deepseek",
  "deepseek": {
    "apiKey": "${DEEPSEEK_API_KEY}",
    "baseUrl": "https://api.deepseek.com",
    "model": "deepseek-v4-flash"
  }
}
```

| Option | Description | Default |
|---|---|---|
| apiKey | Your DeepSeek API key | - |
| baseUrl | DeepSeek API endpoint | https://api.deepseek.com |
| model | DeepSeek model ID | deepseek-v4-flash |

## Models

| Model | Best for |
|---|---|
| deepseek-v4-flash | Fast default model for coding and agent loops |
| deepseek-v4-pro | Stronger reasoning and code review |
| deepseek-chat | Legacy non-thinking alias |
| deepseek-reasoner | Legacy reasoning alias |

## Troubleshooting

| Issue | Fix |
|---|---|
| Invalid API key | Check DEEPSEEK_API_KEY or deepseek.apiKey. |
| Model not found | Use deepseek-v4-flash or deepseek-v4-pro. |
| Network or proxy issues | Set deepseek.baseUrl only when routing through an approved proxy. |