Zum Inhalt

Continue.dev Cheat Sheet

Im Überblick

Continue.dev ist ein Open-Source-KI-Code-Assistent, der Entwickler in volle Kontrolle bringt. Es ist modell-agnostisch, hoch anpassbar und kann mit jedem LLM-Anbieter oder lokal gehosteten Modellen laufen. Perfekt für Teams, die Flexibilität und Privatsphäre benötigen.

ZEITSCHRIFTEN Note: Freie und Open Source, unterstützt mehrere LLM-Anbieter

• Installation

VS Code

```bash

Install from VS Code Marketplace

Search for "Continue" in Extensions

Or install via command line

code --install-extension Continue.continue

Verify installation

Check Extensions panel for Continue

```_

  • JetBrains IDEs ```bash

Install from JetBrains Plugin Repository

Go to File > Settings > Plugins

Search for "Continue" and install

Restart IDE after installation

```_

Manuelle Installation

```bash

Clone repository

git clone https://github.com/continuedev/continue.git cd continue

Install dependencies

npm install

Build extension

npm run build

Package for VS Code

npm run package ```_

Die erste Konfiguration

Basic Setup

json // ~/.continue/config.json \\\\{ "models": [ \\\\{ "title": "GPT-4", "provider": "openai", "model": "gpt-4", "apiKey": "your-openai-api-key" \\\\} ], "tabAutocompleteModel": \\\\{ "title": "Codestral", "provider": "mistral", "model": "codestral-latest", "apiKey": "your-mistral-api-key" \\\\} \\\\}_

Multiple Model Setup

json \\\\{ "models": [ \\\\{ "title": "GPT-4 Turbo", "provider": "openai", "model": "gpt-4-turbo-preview", "apiKey": "your-openai-key" \\\\}, \\\\{ "title": "Claude 3", "provider": "anthropic", "model": "claude-3-opus-20240229", "apiKey": "your-anthropic-key" \\\\}, \\\\{ "title": "Local Llama", "provider": "ollama", "model": "llama2:7b" \\\\} ] \\\\}_

Modellanbieter

OpenAI

json \\\\{ "title": "GPT-4", "provider": "openai", "model": "gpt-4", "apiKey": "sk-...", "apiBase": "https://api.openai.com/v1", "requestOptions": \\\\{ "temperature": 0.7, "maxTokens": 2048 \\\\} \\\\}_

Anthropic Claude

json \\\\{ "title": "Claude 3", "provider": "anthropic", "model": "claude-3-opus-20240229", "apiKey": "sk-ant-...", "requestOptions": \\\\{ "temperature": 0.5, "maxTokens": 4096 \\\\} \\\\}_

Lokale Modelle (Ollama)

json \\\\{ "title": "Local Code Llama", "provider": "ollama", "model": "codellama:7b", "apiBase": "http://localhost:11434", "requestOptions": \\\\{ "temperature": 0.2, "numPredict": 1024 \\\\} \\\\}_

Azure OpenAI

json \\\\{ "title": "Azure GPT-4", "provider": "azure", "model": "gpt-4", "apiKey": "your-azure-key", "apiBase": "https://your-resource.openai.azure.com", "apiVersion": "2023-12-01-preview", "deploymentName": "gpt-4-deployment" \\\\}_

OpenRouter

json \\\\{ "title": "OpenRouter GPT-4", "provider": "openrouter", "model": "openai/gpt-4", "apiKey": "sk-or-...", "requestOptions": \\\\{ "temperature": 0.7 \\\\} \\\\}_

oder Basisnutzung

Chat Schnittstelle

```bash

Open Continue chat

Ctrl+Shift+M (VS Code) Ctrl+Shift+J (JetBrains)

Quick chat

Ctrl+I (VS Code) Ctrl+Shift+I (JetBrains)

Chat with selection

1. Select code

2. Right-click > "Continue: Chat"

3. Or use Ctrl+Shift+M

```_

Code Generation

```javascript // Type comment and use Continue // Generate a REST API endpoint for user authentication // Continue will suggest implementation

// Or use chat: // "Create a React component for file upload with drag and drop" ```_

Code Erklärung

```bash

Select code and ask:

"Explain this function" "What does this regex do?" "How does this algorithm work?" "What are the potential issues with this code?" ```_

Schlüsselanhänger

Shortcut Action IDE
INLINE_CODE_37 Open chat VS Code
INLINE_CODE_38 Quick chat VS Code
INLINE_CODE_39 Select code for chat VS Code
INLINE_CODE_40 Open chat JetBrains
INLINE_CODE_41 Quick chat JetBrains
INLINE_CODE_42 Accept autocomplete All
INLINE_CODE_43 Dismiss autocomplete All
INLINE_CODE_44 Apply suggestion All

Erweiterte Konfiguration

Custom Slash Commands

json \\\\{ "slashCommands": [ \\\\{ "name": "test", "description": "Generate unit tests", "prompt": "Generate comprehensive unit tests for the selected code. Include edge cases and mock dependencies." \\\\}, \\\\{ "name": "optimize", "description": "Optimize code performance", "prompt": "Analyze the selected code and suggest performance optimizations. Focus on time complexity and memory usage." \\\\}, \\\\{ "name": "security", "description": "Security review", "prompt": "Review the selected code for security vulnerabilities. Check for common issues like SQL injection, XSS, and authentication flaws." \\\\} ] \\\\}_

Context Providers

json \\\\{ "contextProviders": [ \\\\{ "name": "codebase", "params": \\\\{ "nRetrieve": 25, "nFinal": 5, "useReranking": true \\\\} \\\\}, \\\\{ "name": "file", "params": \\\\{\\\\} \\\\}, \\\\{ "name": "folder", "params": \\\\{\\\\} \\\\}, \\\\{ "name": "git", "params": \\\\{\\\\} \\\\}, \\\\{ "name": "github", "params": \\\\{ "repos": [ \\\\{ "owner": "microsoft", "repo": "vscode" \\\\} ] \\\\} \\\\} ] \\\\}_

Custom Model Configuration

json \\\\{ "models": [ \\\\{ "title": "Custom Local Model", "provider": "openai", "model": "custom-model", "apiBase": "http://localhost:8000/v1", "apiKey": "not-needed", "requestOptions": \\\\{ "temperature": 0.3, "maxTokens": 2048, "stop": ["<|endoftext|>"] \\\\} \\\\} ] \\\\}_

Autocomplete Konfiguration

Tab Autocomplete

json \\\\{ "tabAutocompleteModel": \\\\{ "title": "Codestral", "provider": "mistral", "model": "codestral-latest", "apiKey": "your-mistral-key" \\\\}, "tabAutocompleteOptions": \\\\{ "useCopyBuffer": true, "maxPromptTokens": 1024, "prefixPercentage": 0.85, "maxSuffixPercentage": 0.25, "debounceDelay": 300 \\\\} \\\\}_

Autocomplete Provider

json \\\\{ "tabAutocompleteModel": \\\\{ "title": "Local Autocomplete", "provider": "ollama", "model": "deepseek-coder:6.7b", "requestOptions": \\\\{ "temperature": 0.1, "numPredict": 256 \\\\} \\\\} \\\\}_

Im Kontext Management

Dateikontext

```bash

Add files to context

@file:src/utils/auth.js

Add multiple files

@file:src/components/Button.tsx @file:src/styles/button.css

Add entire folders

@folder:src/components ```_

Codebase Context

```bash

Search codebase for relevant context

@codebase "authentication functions"

Find similar code patterns

@codebase "React hooks for API calls"

Search for specific implementations

@codebase "error handling middleware" ```_

Git Context

```bash

Reference git history

@git "recent changes to authentication"

Compare branches

@git "differences between main and feature-branch"

Reference specific commits

@git "changes in commit abc123" ```_

Individuelle Integrationen

Datenbankintegration

json \\\\{ "contextProviders": [ \\\\{ "name": "database", "params": \\\\{ "connectionString": "postgresql://user:pass@localhost:5432/db", "tables": ["users", "orders", "products"] \\\\} \\\\} ] \\\\}_

API Dokumentation

json \\\\{ "contextProviders": [ \\\\{ "name": "docs", "params": \\\\{ "urls": [ "https://docs.stripe.com/api", "https://docs.github.com/en/rest" ] \\\\} \\\\} ] \\\\}_

Jira Integration

json \\\\{ "contextProviders": [ \\\\{ "name": "jira", "params": \\\\{ "domain": "your-company.atlassian.net", "token": "your-jira-token", "email": "your-email@company.com" \\\\} \\\\} ] \\\\}_

Sprache-spezifische Konfiguration

Python Setup

json \\\\{ "models": [ \\\\{ "title": "Python Specialist", "provider": "openai", "model": "gpt-4", "systemMessage": "You are a Python expert. Always follow PEP 8 style guidelines and use type hints." \\\\} ] \\\\}_

JavaScript/TypScript

json \\\\{ "models": [ \\\\{ "title": "TS Expert", "provider": "anthropic", "model": "claude-3-opus-20240229", "systemMessage": "You are a TypeScript expert. Always use strict typing and modern ES6+ features." \\\\} ] \\\\}_

Rust Configuration

json \\\\{ "models": [ \\\\{ "title": "Rust Helper", "provider": "ollama", "model": "codellama:7b", "systemMessage": "You are a Rust expert. Focus on memory safety, performance, and idiomatic Rust code." \\\\} ] \\\\}_

Team Konfiguration

Geteilte Konfiguration

json // .continue/config.json (in project root) \\\\{ "models": [ \\\\{ "title": "Team GPT-4", "provider": "openai", "model": "gpt-4", "apiKey": "$\\\\{OPENAI_API_KEY\\\\}" \\\\} ], "slashCommands": [ \\\\{ "name": "review", "description": "Code review following team standards", "prompt": "Review this code according to our team's coding standards: $\\\\{TEAM_STANDARDS\\\\}" \\\\} ] \\\\}_

Umgebungsvariablen

```bash

.env file

OPENAI_API_KEY=sk-... ANTHROPIC_API_KEY=sk-ant-... TEAM_STANDARDS="Use TypeScript, follow ESLint rules, include unit tests"

Use in config

\\{ "apiKey": "$\\{OPENAI_API_KEY\\}" \\} ```_

 Leistungsoptimierung

Caching Configuration

json \\\\{ "embeddingsProvider": \\\\{ "provider": "openai", "model": "text-embedding-ada-002", "apiKey": "your-key" \\\\}, "reranker": \\\\{ "name": "cohere", "params": \\\\{ "apiKey": "your-cohere-key", "model": "rerank-english-v2.0" \\\\} \\\\} \\\\}_

Lokale Einbettungen

json \\\\{ "embeddingsProvider": \\\\{ "provider": "ollama", "model": "nomic-embed-text", "apiBase": "http://localhost:11434" \\\\} \\\\}_

Fehlerbehebung

Häufige Fragen

```bash

Extension not loading

1. Check VS Code/JetBrains version compatibility

2. Restart IDE

3. Reinstall extension

4. Check Continue logs

API key issues

1. Verify API key format

2. Check API key permissions

3. Test API key with curl

4. Check rate limits

Model not responding

1. Check internet connection

2. Verify model name

3. Check API endpoint

4. Review request options

```_

Debug Mode

json \\\\{ "allowAnonymousTelemetry": false, "logLevel": "debug" \\\\}_

Log Analysis

```bash

VS Code logs location

# Windows: %APPDATA%\Code\logs\

macOS: ~/Library/Logs/Code/

Linux: ~/.config/Code/logs/

JetBrains logs

Check IDE logs directory

Help > Show Log in Explorer/Finder

```_

oder Best Practices

Effektive Promption

```javascript // ❌ Vague request "Fix this code"

// ✅ Specific request "Optimize this function for better performance and add error handling for edge cases"

// ❌ No context "Create a component"

// ✅ With context "Create a React component for displaying user profiles with TypeScript interfaces and proper prop validation" ```_

Context Management

```bash

Use relevant context providers

@codebase "similar authentication patterns" @file:src/types/user.ts

Be specific about requirements

"Using the User interface from @file:src/types/user.ts, create a validation function" ```_

Modellauswahl

```bash

Use appropriate models for tasks

- GPT-4: Complex reasoning, architecture decisions

- Claude: Long context, detailed explanations

- Local models: Privacy, offline work

- Specialized models: Domain-specific tasks

```_

Ressourcen