WiseGPT (Coding Assistant by DhiWise)

Analyzes the entire codebase within a VS Code workspace to build a semantic understanding of code patterns, architecture, and style conventions. The extension sends codebase metadata to DhiWise backend servers which index and vectorize the code for context-aware generation. Uses @codebase mention syntax in chat to trigger full repository context retrieval, enabling the AI to understand existing patterns, naming conventions, and architectural decisions before generating new code.

Integrates with task management platforms (Jira, Trello, Asana, ClickUp) to extract task descriptions and requirements, then generates production-ready code that implements those tasks. The extension reads task metadata including title, description, acceptance criteria, and linked resources, sends them to the DhiWise backend along with codebase context, and returns generated code that matches the project's existing style and architecture. Eliminates the need for manual prompt engineering by converting structured task data into code generation requests.

Generates code across multiple programming languages and frameworks, with support claimed for 'all programming languages and frameworks'. The extension analyzes the project's technology stack and generates code in the appropriate language and framework, using language-specific idioms and best practices. Backend inference adapts to language-specific patterns, syntax, and conventions, ensuring generated code is idiomatic rather than generic translations.

Operates on a freemium pricing model with free tier access to basic code generation and chat features, while advanced features like vulnerability detection and code implementation for tasks are restricted to enterprise users. The extension manages feature access through backend authentication and account tier checking, enabling free users to access core capabilities while reserving advanced security and automation features for paid tiers. Specific free tier limits (requests per day, codebase size, etc.) are not documented.

Converts Figma design files into functional code by analyzing design components, layouts, and styling, then generates code using the project's existing UI libraries and component patterns. The extension reads Figma design metadata (components, constraints, colors, typography) and sends it to the DhiWise backend along with codebase context, which then generates code that reuses existing project components and styling conventions rather than creating new ones. Supports integration with DhiWise Design Converter projects to pull source code directly into the IDE.

Provides real-time code completion suggestions as developers type, with suggestions personalized to match the project's coding style and patterns. The extension monitors editor changes and sends partial code context to the DhiWise backend, which returns completion suggestions that align with existing code conventions, naming patterns, and architectural decisions. Supports both traditional autocompletion and comment-based code generation where developers write comments describing desired functionality and the AI generates matching code.

Scans code for security vulnerabilities and generates fixes that remediate identified issues while maintaining code functionality. The extension analyzes the codebase for common vulnerability patterns (SQL injection, XSS, insecure dependencies, etc.) and sends findings to the DhiWise backend, which generates corrected code that fixes the vulnerability using secure coding practices appropriate to the project's technology stack. Integrates with the codebase context to ensure fixes follow existing patterns and conventions.

Automatically generates unit tests, integration tests, and test cases based on code implementation and task requirements. The extension analyzes function signatures, logic flow, and acceptance criteria from linked tasks, then generates test code that covers normal cases, edge cases, and error conditions. Generated tests follow the project's testing framework conventions and style, integrating with existing test suites rather than creating isolated test files.

Automatically generates documentation for code including function documentation, API documentation, README sections, and inline comments. The extension analyzes code structure, function signatures, parameters, return types, and business logic, then generates documentation that explains purpose, usage, parameters, and examples. Documentation is generated in formats matching project conventions (JSDoc, Sphinx, Markdown, etc.) and includes code examples derived from existing usage patterns in the codebase.

Analyzes code for performance improvements, readability enhancements, and refactoring opportunities, then generates optimized versions of the code. The extension identifies inefficient patterns, redundant code, and opportunities for abstraction, sends them to the DhiWise backend for analysis, and generates refactored code that improves performance or maintainability while preserving functionality. Refactoring suggestions respect project conventions and avoid introducing breaking changes.

Provides a conversational interface for asking questions about the codebase, understanding code behavior, and navigating complex logic. The extension maintains a chat context where developers can ask natural language questions about code functionality, architecture, or specific implementations, and the AI responds with explanations, code references, and navigation suggestions. Uses @codebase mention syntax to explicitly include repository context in questions, enabling precise answers about specific code patterns or architectural decisions.

Allows developers to attach files and images to chat messages to provide additional context for code generation requests. The extension supports attaching code files, design images, configuration files, and other artifacts that provide context for generation tasks. Attached content is sent to the DhiWise backend along with the chat message, enabling the AI to generate code that incorporates specific requirements from attached files or design mockups without requiring manual context description.