Scitely AI Agent for VS Code
A powerful AI coding assistant extension for Visual Studio Code, powered by Scitely's multi-model AI platform.
Features
Multi-Model AI Support
- DeepSeek: V3.2, R1, V3 - Advanced reasoning and coding models
- Qwen3: Max, Coder Plus, VL Plus - High-performance language and vision models
- Kimi K2: Fast and efficient chat model
- GLM 4.6: Versatile language model
- Custom Providers: OpenAI-compatible API support
Agent Modes
- General Mode: Answer questions, explain concepts, brainstorm ideas
- Coding Mode: Write, debug, refactor, and review code
- Vision Mode: Analyze images, screenshots, diagrams, and UI designs
Smart Features
- Auto Model Selection: Automatically picks the best model for your task
- Context Awareness: Uses your current file and selection for better responses
- Code Actions: Quick actions for explain, fix, refactor, and test generation
- Image Support: Paste or upload images for vision analysis
- Streaming Responses: Real-time response streaming for faster feedback
Editor Integration
- Right-click context menu for quick code actions
- Keyboard shortcuts for common operations
- Insert generated code directly into editor
- Copy code blocks with one click
Installation
- Install from VS Code Marketplace or download the
.vsix file
- Open VS Code and go to Extensions (Ctrl+Shift+X)
- Search for "Scitely AI Agent" and click Install
Getting Started
Sign Up / Login
- Open the Scitely sidebar (click the Scitely icon in the activity bar)
- Click "Create Account" to sign up, or "Sign In" if you have an account
- Enter your email and password
Using the Agent
- Open the Scitely panel from the sidebar
- Type your question or request in the input box
- Press Enter or click Send
Quick Actions
Select code in the editor and:
- Right-click → "Scitely: Explain Code"
- Right-click → "Scitely: Fix Code"
- Right-click → "Scitely: Generate Tests"
- Or use Ctrl+Shift+S to send selection to chat
Vision Mode
- Switch to Vision mode in the dropdown
- Click the image icon or paste an image (Ctrl+V)
- Ask questions about the image
Keyboard Shortcuts
| Action |
Windows/Linux |
macOS |
| Open Agent |
Ctrl+Shift+A |
Cmd+Shift+A |
| Send Selection |
Ctrl+Shift+S |
Cmd+Shift+S |
Configuration
Open Settings (Ctrl+,) and search for "Scitely":
| Setting |
Description |
Default |
scitely.defaultProvider |
Default AI provider |
scitely |
scitely.defaultModel |
Default model (or 'auto') |
auto |
scitely.customApiBaseUrl |
Custom API base URL |
- |
scitely.autoContext |
Include current file context |
true |
scitely.maxContextLines |
Max context lines |
500 |
scitely.streamResponses |
Enable streaming |
true |
Using Custom Providers
OpenAI
- Go to Settings → Scitely
- Set
defaultProvider to openai
- Use Command Palette → "Scitely: Login" to configure API key
Custom OpenAI-Compatible API
- Set
scitely.customApiBaseUrl to your API endpoint
- Set
defaultProvider to custom
- Configure your API key via the extension
Privacy & Security
- Secure Storage: All API keys and tokens are stored in VS Code's SecretStorage
- No Telemetry: We don't collect or send any usage data
- Local Processing: Code context is processed locally before sending
- JWT Auth: Secure authentication with token-based auth
Troubleshooting
"Authentication required"
- Make sure you're logged in
- Try logging out and back in
"Model not available"
- Check your API key is correct
- Verify the model is available for your plan
Slow Responses
- Try disabling streaming in settings
- Use a smaller model for faster responses
Support
License
MIT License - see LICENSE for details.
Made with AI by Scitely
| |
