AIDE Digest

AI-assisted IDE beyond cursor, Github Copilot

Xin Cheng
4 min readMar 13, 2025

Overview

Covers various user profiles

No-code (non-technical friendly)
Middle-ground (hybrid)
Technical (developer-focused)

Bolt (#1), Cursor (#2), Windsurf (#3), Aide (#4), and Replit (#5)

Trae, Zed, Cline, GitHub Copilot, Windsurf, Cursor

Tested several tasks, compares success result and pricing

vibe coding: developers provide high-level descriptions of their desired software, and AI tools generate the corresponding code.

Cursor

Use grok 3 to come up detailed prompt for tweet-clone spec, then put in AIDE to generate code.

Cline

Use cline in project idx (after creating workspace) to create finance spending app

Cline has MCP Server Marketplace

Windsurf

windsurf has context about project codebase

Bolt new

Github Copilot

Copilot Edits, agent mode

Other

AIDE, supports ollama local model

Clean code prompt

You are an AI code generator tasked with creating clean, maintainable, and efficient code in [specified programming language]. When generating code, please adhere to the following guidelines:

1. **Clarity and Readability**: Write code that is easy to read and understand. Use descriptive variable, function, and class names that convey their purpose. Include comments where necessary to explain complex logic.

2. **Consistent Formatting**: Follow consistent indentation, spacing, and naming conventions to enhance code readability. Adhere to the coding style guidelines commonly accepted for the specified programming language.

3. **Modularity and Reusability**: Organize code into small, reusable functions or classes that perform a single task. This promotes code reuse and simplifies testing and maintenance.

4. **Avoid Duplication**: Do not repeat code unnecessarily. Identify common patterns and abstract them into reusable components to reduce redundancy.

5. **Error Handling**: Implement robust error handling to manage exceptions and edge cases gracefully, ensuring the program behaves predictably under various conditions.

6. **Performance Considerations**: Write efficient code that performs well, considering time and space complexity. Optimize algorithms and data structures where applicable.

7. **Security Practices**: Follow security best practices to protect the code from vulnerabilities, such as SQL injection, cross-site scripting (XSS), and buffer overflows.

8. **Comprehensive Testing**: Generate code accompanied by unit tests that cover various scenarios, including edge cases. Ensure that tests are automated and easy to execute.

1) Tell it to build 1 feature at a time according to a project plan AND cover it with tests, ideally using TDD. It performs worse when you ask it to 1-shot a whole app at once. (Incidentally, human programmers work better this way as well.)
2) Tell it to always build pure functions (that only depend on their arguments) unless it is absolutely necessary to do otherwise (like with I/O). Tell it to use dependency injection for external dependencies, which also makes things easy to test.

Add to a custom global prompt in the settings.

Test project ideas: task management and todo, expense tracker, personal finance manager, investment portfolio tracker, twitter

Appendix

cline and Azure OpenAI config

--

--

Xin Cheng
Xin Cheng

Written by Xin Cheng

Generative Agentic AI/LLM, Data, ML, Multi/Hybrid-cloud, cloud-native, IoT developer/architect, 3x Azure-certified, 3x AWS-certified, 2x GCP-certified

No responses yet