Home›Third-party Integrations

Configure DeepSeek API for Aider, Continue, and Cursor

DeepSeek provides native OpenAI API protocol compatibility. Connect DeepSeek models to Aider, Continue, and Cursor with standard endpoint configuration.

Maintained by DeepSeekDSH (independent site)Updated: 2026-09-13

On this page

OpenAI-compatible parameters

Base URL: https://api.deepseek.com or https://api.deepseek.com/v1
Chat and coding model: deepseek-chat (DeepSeek-V3 for fast generation and autocompletion)
Reasoning model: deepseek-reasoner (DeepSeek-R1 for complex problem solving)
API key: Generated in your DeepSeek Platform dashboard.

TOOL 1

Continue extension setup (VS Code & JetBrains)

In Continue configuration (~/.continue/config.yaml or config.json), add DeepSeek models:

models:
  - name: "DeepSeek Chat"
    provider: "openai"
    model: "deepseek-chat"
    apiKey: "sk-your_deepseek_api_key"
    apiBase: "https://api.deepseek.com/v1"
  - name: "DeepSeek Reasoner (R1)"
    provider: "openai"
    model: "deepseek-reasoner"
    apiKey: "sk-your_deepseek_api_key"
    apiBase: "https://api.deepseek.com/v1"

tabAutocompleteModel:
  name: "DeepSeek Coder Autocomplete"
  provider: "openai"
  model: "deepseek-chat"
  apiKey: "sk-your_deepseek_api_key"
  apiBase: "https://api.deepseek.com/v1"

TOOL 2

Aider CLI pair programming setup

Aider has built-in DeepSeek support. Export the key and launch:

# Native DeepSeek provider (recommended)
export DEEPSEEK_API_KEY="sk-your_deepseek_api_key"
aider --model deepseek/deepseek-chat

# Generic OpenAI-compatible variables:
# export OPENAI_API_BASE="https://api.deepseek.com"
# export OPENAI_API_KEY="sk-your_deepseek_api_key"

Or launch with the reasoning model:

aider --model deepseek/deepseek-reasoner

TOOL 3

Cursor editor setup

In Cursor Settings → Models:

Enable OpenAI API Key and paste your key.
Check Override OpenAI Base URL and enter https://api.deepseek.com/v1.
Add and enable deepseek-chat in the model list.

Tip: If streaming reasoning tokens cause syntax glitches in your Cursor version, use deepseek-chat for standard code generation.

Prompt caching and latency efficiency

DeepSeek API automatically caches prompt prefixes on the server. In multi-turn coding sessions, cached tokens cost significantly less and return responses faster.

Troubleshooting

404 Model Not Found

Ensure the model ID matches deepseek-chat or deepseek-reasoner.

401 Invalid API Key

Check that the key begins with sk- with no extra spaces.