DeepSeek Chat Pro
Native VS Code language model provider for DeepSeek v4. Optimized for high-speed reasoning and complex workspace tasks.
Key Features
- DeepSeek v4 Native: Full support for
v4-flash and v4-pro architectures.
- Granular Reasoning: Select from
thinking-high, thinking-max, or thinking-off variants in the model picker.
- KV Cache Hardening: Deterministic history management for faster response times and lower latency via prefix caching.
- Workspace Aware: Built-in
read_file tool support for direct code context integration.
- Secure Storage: API keys are stored encrypted via VS Code SecretStorage.
Getting Started
- Install the extension.
- Run the command DeepSeek: Set API Key from the Command Palette.
- Open the Chat view and select a DeepSeek model variant.
- Start coding with DeepSeek v4 reasoning.
Extension Settings
| Setting |
Description |
deepseekChat.baseUrl |
Customize the API endpoint (defaults to DeepSeek official). |
deepseekChat.selectedModel |
Choose your preferred default model. |
License
MIT
| |
