Continue.dev Setup Guide: Context Configuration That Actually Works

{ "models": [ { "title": "Claude 3.5 Sonnet (Primary)", "provider": "anthropic", "model": "claude-3-5-sonnet-20241022", "apiKey": "your-anthropic-api-key", "contextLength": 200000, "completionOptions": { "temperature": 0.1, "maxTokens": 2048 } }, { "title": "GPT-4 (Fallback)", "provider": "openai", "model": "gpt-4-turbo", "apiKey": "your-openai-api-key", "contextLength": 128000 } ], "tabAutocompleteModel": { "title": "Claude 3.5 Sonnet", "provider": "anthropic", "model": "claude-3-5-sonnet-20241022", "apiKey": "your-anthropic-api-key" }, "customCommands": [ { "name": "review", "prompt": "Review this code for:\n1. Adherence to our coding standards (see docs/CODING_STANDARDS.md)\n2. Performance implications\n3. Security vulnerabilities\n4. Test coverage requirements\n5. Documentation needs\n\nProvide specific, actionable feedback." }, { "name": "test", "prompt": "Generate comprehensive tests for this code following our testing patterns in tests/ directory. Include:\n1. Unit tests covering all branches\n2. Edge cases and error scenarios\n3. Integration tests if needed\n4. Mock setup following our mocking conventions" }, { "name": "docs", "prompt": "Generate documentation for this code following our documentation standards. Include:\n1. Function/class purpose and behavior\n2. Parameter descriptions and types\n3. Return value documentation\n4. Usage examples\n5. Error conditions and exceptions" } ], "contextProviders": [ { "name": "codebase", "params": { "nResults": 30, "nRetrieve": 100 } }, { "name": "diff", "params": {} }, { "name": "open", "params": {} }, { "name": "terminal", "params": {} }, { "name": "problems", "params": {} }, { "name": "folder", "params": { "folders": [ "src/", "lib/", "components/", "services/", "utils/", "types/", "docs/" ] } }, { "name": "docs", "params": { "startUrl": "https://docs.yourproject.com" } } ], "allowAnonymousTelemetry": false, "embeddingsProvider": { "provider": "openai", "model": "text-embedding-3-small", "apiKey": "your-openai-api-key" } }

# .continueignore (project root) # Exclude files that add noise to context # Dependencies node_modules/ vendor/ .venv/ venv/ __pycache__/ # Build artifacts dist/ build/ target/ .next/ .nuxt/ # Logs and temporary files *.log *.tmp .DS_Store Thumbs.db # IDE files .vscode/settings.json .idea/ # Large data files *.csv *.json *.sql *.dump # Generated code (unless you want AI to learn from it) generated/ auto-generated/ # Test fixtures with large data test/fixtures/large-data/

# Create .continue/system-prompt.md in your project root You are an expert software engineer working on [PROJECT_NAME], a [PROJECT_DESCRIPTION]. ## Codebase Context - **Architecture**: [Microservices/Monolith/Serverless/etc] - **Primary Language**: [TypeScript/Python/Java/etc] - **Framework**: [Next.js/Django/Spring Boot/etc] - **Database**: [PostgreSQL/MongoDB/Redis/etc] - **Cloud Platform**: [AWS/GCP/Azure/etc] ## Coding Standards 1. **Naming Conventions**: - Variables: camelCase for JS/TS, snake_case for Python - Functions: verb-noun pattern (getUserById, calculateTotalPrice) - Classes: PascalCase with descriptive names - Constants: UPPER_SNAKE_CASE 2. **Code Organization**: - Follow feature-based folder structure - Separate concerns: models, services, controllers, utils - Keep functions small and single-purpose - Use dependency injection for testability 3. **Error Handling**: - Always handle errors explicitly - Use custom error classes for business logic errors - Log errors with appropriate context - Never expose internal errors to users 4. **Testing Requirements**: - Minimum 80% code coverage - Test edge cases and error scenarios - Use factory functions for test data - Mock external dependencies ## Business Domain Knowledge [Include specific business rules, domain concepts, and terminology] ## Architecture Patterns [Document your specific patterns: repository pattern, service layer, etc] ## Security Requirements [List security considerations: authentication, authorization, data validation] When generating code: 1. Follow established patterns in the codebase 2. Use meaningful variable and function names based on business context 3. Include appropriate error handling and logging 4. Consider performance and security implications 5. Generate accompanying tests when appropriate

# Recommended docs/ folder structure for Continue.dev context docs/ ├── README.md # Project overview and setup ├── ARCHITECTURE.md # System architecture and design decisions ├── CODING_STANDARDS.md # Detailed coding conventions ├── API_DESIGN.md # API design patterns and conventions ├── DEPLOYMENT.md # Deployment processes and requirements ├── TROUBLESHOOTING.md # Common issues and solutions ├── examples/ │ ├── service-patterns.md # Service layer examples │ ├── api-endpoints.md # API endpoint examples │ ├── database-queries.md # Database interaction examples │ └── error-handling.md # Error handling examples └── decisions/ ├── 001-database-choice.md ├── 002-authentication-strategy.md └── 003-caching-approach.md

# docs/examples/service-patterns.md ## Service Layer Patterns ### User Service Example ```typescript export class UserService { constructor( private userRepository: UserRepository, private logger: Logger, private eventBus: EventBus ) {} async createUser(userData: CreateUserRequest): Promise { try { // Validation const validatedData = await this.validateUserData(userData); // Business logic const user = await this.userRepository.create(validatedData); // Events await this.eventBus.publish(new UserCreatedEvent(user)); // Logging this.logger.info('User created successfully', { userId: user.id }); return user; } catch (error) { this.logger.error('Failed to create user', { error, userData }); throw new UserCreationError('Unable to create user', error); } } } ``` ### Database Query Patterns ```typescript export class UserRepository { async findActiveUsersByRole(role: UserRole): Promise { return this.db.query( `SELECT * FROM users WHERE role = $1 AND status = 'active' ORDER BY created_at DESC`, [role] ); } } ``` ### Error Handling Patterns ```typescript export class BusinessLogicError extends Error { constructor( message: string, public code: string, public statusCode: number = 400 ) { super(message); this.name = 'BusinessLogicError'; } } ```

# Add to config.json customCommands array { "name": "refactor", "prompt": "Refactor this code following our established patterns:\n1. Extract reusable functions\n2. Improve naming for clarity\n3. Add proper error handling\n4. Ensure single responsibility principle\n5. Add JSDoc comments\n\nMaintain the same functionality while improving code quality." }, { "name": "optimize", "prompt": "Optimize this code for performance:\n1. Identify potential bottlenecks\n2. Suggest algorithmic improvements\n3. Optimize database queries if present\n4. Reduce memory usage where possible\n5. Maintain readability\n\nExplain the optimizations made." }, { "name": "security", "prompt": "Review this code for security vulnerabilities:\n1. Input validation issues\n2. SQL injection risks\n3. XSS vulnerabilities\n4. Authentication/authorization gaps\n5. Data exposure risks\n\nProvide specific fixes for any issues found." }, { "name": "api", "prompt": "Generate a complete API endpoint following our patterns:\n1. Input validation with proper types\n2. Authentication middleware\n3. Business logic in service layer\n4. Proper error responses\n5. OpenAPI documentation\n6. Unit tests\n\nInclude all necessary imports and dependencies." }

# Custom context provider for team knowledge { "name": "teamKnowledge", "params": { "sources": [ { "type": "confluence", "url": "https://company.atlassian.net/wiki", "spaces": ["DEV", "ARCH", "API"] }, { "type": "notion", "database_ids": ["your-coding-standards-db-id"] }, { "type": "github", "repos": ["company/shared-libraries", "company/api-templates"], "include_issues": true, "include_prs": true } ] } }

# Optimize context usage for better performance { "contextProviders": [ { "name": "codebase", "params": { "nResults": 15, // Reduced from default 30 "nRetrieve": 50, // Reduced from default 100 "useGitIgnore": true, // Respect .gitignore "maxFileSize": 50000 // Skip large files } } ], "reranker": { "name": "cohere", "params": { "model": "rerank-english-v2.0", "apiKey": "your-cohere-api-key" } } }

{ "embeddingsProvider": { "provider": "openai", "model": "text-embedding-3-small", "apiKey": "your-openai-api-key", "cache": { "enabled": true, "maxCacheSize": "1GB", "ttl": 86400 // 24 hours } } }

# Week 1: Setup and Configuration - Install Continue.dev for senior developers - Configure basic context providers - Create initial documentation - Test with simple code generation tasks # Week 2: Pattern Training - Add team-specific system prompts - Create custom slash commands - Document successful patterns - Train AI on existing codebase patterns # Week 3: Team Training - Onboard remaining team members - Conduct training sessions - Establish best practices - Create feedback collection process # Week 4: Optimization - Analyze usage patterns - Optimize context providers - Refine custom commands - Measure productivity improvements

# Track Continue.dev effectiveness productivity_metrics = { "code_generation_speed": "time_to_generate_working_code", "revision_rate": "percentage_of_ai_code_that_needs_modification", "pattern_consistency": "adherence_to_team_coding_standards", "test_coverage": "test_coverage_of_ai_generated_code", "bug_rate": "bugs_introduced_by_ai_generated_code" } # Target improvements after 30 days: target_improvements = { "code_generation_speed": "+67%", "revision_rate": "-45%", "pattern_consistency": "+89%", "test_coverage": "maintained_or_improved", "bug_rate": "no_increase_vs_manually_written_code" }