[GitHub Global] Translate Vibe Coding 零基础教程/10 编程工具/09 我的 AI 工具箱推荐.md to en
This commit is contained in:
yupi-translate-app[bot]
committed by
GitHub
GitHub
parent
2cde17cab4
commit
1e5d66db3b
@@ -0,0 +1,301 @@
|
||||
# My AI Toolbox Recommendations
|
||||
|
||||
> My Top 20 Most Used AI Products This Year
|
||||
|
||||
Hello, I'm Yupi, a former full-stack developer at Tencent and an [AI programming blogger](https://space.bilibili.com/12890453) with over 2 million followers across platforms. I'm also the creator of [AI Navigation](https://ai.codefather.cn) and [Programming Navigation](https://www.codefather.cn), among more than 10 self-developed products.
|
||||
|
||||
In previous articles, we systematically explored various types of AI programming tools: from AI model selection, no-code platforms, AI code editors, command-line tools, IDE plugins, AI agent platforms, to auxiliary toolkits.
|
||||
|
||||
So, the questions arise:
|
||||
|
||||
- Among so many tools, which ones do you actually use, Yupi?
|
||||
- How do you combine these tools in actual work?
|
||||
- Are there any practical experiences you can share?
|
||||
|
||||
In this article, I'll share my Top 20 most-used AI products in 2025, along with my practical insights, hoping to provide you with some reference.
|
||||
|
||||
## 1. AI Programming Tools
|
||||
|
||||
As a programmer, AI programming tools are what I use the most. Below are my main programming tools this year.
|
||||
|
||||
### Cursor
|
||||
|
||||
[Cursor](https://cursor.com/) is, in my opinion, the top AI editor in terms of development experience and AI engineering capabilities. I rely on it for major projects, whether it's writing, modifying, or refactoring code—Cursor's Agent mode handles it all.
|
||||
|
||||
I particularly love its Tab completion feature. While coding, the AI predicts what I'm about to write, and pressing Tab accepts the suggestion. It feels like having an experienced driver sitting beside you, finishing your thoughts before you even type them.
|
||||
|
||||
Moreover, Cursor allows seamless switching between multiple models like Claude, GPT, and Gemini, and supports using your own API Key, enabling me to choose the most suitable model for different tasks.
|
||||
|
||||
However, Cursor's biggest drawback is its steep price. I once shared my bill, which exceeded 10,000 yuan in just over a month. So now, I combine it with some free tools to avoid budget explosions.
|
||||
|
||||
|
||||
|
||||
### Claude Code
|
||||
|
||||
[Claude Code](https://claude.ai/code) is my most-used command-line AI programming tool. Its automation level is exceptionally high—it not only generates code but also executes commands, installs dependencies, and creates files automatically.
|
||||
|
||||
I typically use it for simple tasks or quickly setting up project frameworks. For example, if I want to create an image compression tool, I just tell the AI in the command line, and it sets up the entire project for me.
|
||||
|
||||
|
||||
|
||||
Additionally, Claude Code supports Skills, allowing me to write common norms and processes as Skills for the AI to follow automatically, which is particularly useful for avoiding repetitive work and team collaboration.
|
||||
|
||||
Since Claude Code has limitations in China, I often pair it with domestic AI models like Zhipu GLM. This ensures normal usage while keeping costs lower.
|
||||
|
||||
### Cline
|
||||
|
||||
[Cline](https://cline.bot/) is an AI IDE plugin and one of my backups. When Cursor's budget runs out, or I want to develop in VS Code, I turn to Cline.
|
||||
|
||||
|
||||
|
||||
Cline's biggest advantage is that it's completely free and quite comprehensive, achieving about 80% of what Cursor does. Although the experience isn't as smooth as Cursor's, considering it's free, it's already impressive.
|
||||
|
||||
Moreover, Cline supports multiple platforms, not just VS Code but also JetBrains IDEs. This means I can use the same tool across different editors, maintaining a consistent experience.
|
||||
|
||||
## 2. AI Multimedia Creation Tools
|
||||
|
||||
Besides programming, I also use AI for content creation. Below are my most-used multimedia creation tools.
|
||||
|
||||
### Nano Banana
|
||||
|
||||
Nano Banana is this year's most explosive AI image generation tool. I use it to create comic-style knowledge-explanation images and article cover images.
|
||||
|
||||
It has excellent support for Chinese, unlike Midjourney, which requires English prompts. Moreover, the generated images have a consistent style, making them perfect for tutorial illustrations.
|
||||
|
||||
The anime-style images you see here are all created with Nano Banana:
|
||||
|
||||
|
||||
|
||||
Since using Nano Banana, I've completely bid farewell to Midjourney.
|
||||
|
||||
### Jimeng AI
|
||||
|
||||
[Jimeng AI](https://jimeng.jianying.com/) is ByteDance's all-in-one AI creation platform, supporting text-to-image, image-to-video, and more.
|
||||
|
||||
I mainly use it to generate video background scenes. For example, when creating tutorial videos, I use Jimeng AI to generate background visuals. Its video generation speed is fast, and the quality is decent.
|
||||
|
||||
Moreover, Jimeng AI is a domestic platform, offering fast access and no worries about being blocked.
|
||||
|
||||
### Veo 3
|
||||
|
||||
[Veo 3](https://aistudio.google.com/models/veo-3) is Google's most advanced AI video generation model. I use it to create creative short videos.
|
||||
|
||||
Veo 3's video quality is indeed high, with realistic physics and lighting effects. However, it's prohibitively expensive, so I only use it when I need high-quality videos.
|
||||
|
||||
|
||||
|
||||
### Eleven Labs
|
||||
|
||||
[Eleven Labs](https://elevenlabs.io/) is an AI voice synthesis tool known for its incredibly realistic voice imitation.
|
||||
|
||||
Eleven Labs supports multiple languages and voice styles, allowing me to choose different voices for various scenarios. I use it to synthesize voiceovers for creative videos.
|
||||
|
||||
|
||||
|
||||
## 3. AI Large Models
|
||||
|
||||
AI large models are the foundation of all AI tools. Below are the large models I've used the most this year.
|
||||
|
||||
### Zhipu GLM
|
||||
|
||||
[Zhipu GLM](https://bigmodel.cn/) is my most-used domestic large model, excelling in programming and Chinese comprehension.
|
||||
|
||||
I mainly use it with Claude Code to generate non-commercial full-stack projects. GLM's API is affordable and fast, making it ideal for small projects or quick idea validation.
|
||||
|
||||
Zhipu's development has been rapid, with significant improvements in programming capabilities since GLM-4.6, even surpassing Claude Sonnet in some benchmarks! Its IPO was well-deserved.
|
||||
|
||||
### Alibaba Tongyi Qianwen
|
||||
|
||||
Tongyi Qianwen is the large model integrated into most of our team's products. Its advantages include fast generation speed, easy API integration, and debugging via [Alibaba Cloud Bailian Platform](https://bailian.console.aliyun.com/).
|
||||
|
||||
|
||||
|
||||
Our products use it to generate simple content or perform intent recognition. For example, when a user inputs a sentence, we use Tongyi Qianwen to determine the user's intent and call the corresponding function.
|
||||
|
||||
Moreover, Tongyi Qianwen offers free quotas, which are sufficient for small projects.
|
||||
|
||||
### DeepSeek
|
||||
|
||||
[DeepSeek](https://www.deepseek.com/) was once a hit, completely free and strong in programming. However, its product ecosystem and new model release speed have recently lagged.
|
||||
|
||||
I now mainly use it to verify technical knowledge. For example, when writing articles, if I'm unsure about a technical point, I ask DeepSeek. Since it's free, I can ask as much as I want without worrying about costs.
|
||||
|
||||
|
||||
|
||||
### Gemini
|
||||
|
||||
[Gemini](https://gemini.google.com/) boasts powerful reasoning and multimodal understanding capabilities, supporting ultra-long contexts (1 million tokens).
|
||||
|
||||
I mainly use it to analyze large projects. For example, when I take on a project with hundreds of files, I use Gemini to quickly understand the project structure.
|
||||
|
||||
I also use Gemini's web version to gather foreign technology and product research, as it provides professional reports.
|
||||
|
||||
|
||||
|
||||
However, Gemini has certain usage barriers in China, as you know.
|
||||
|
||||
## 4. AI Development Frameworks
|
||||
|
||||
Besides directly using AI tools, I also use some AI development frameworks to build AI applications.
|
||||
|
||||
### Spring AI
|
||||
|
||||
[Spring AI](https://spring.io/projects/spring-ai) is the standard for Java AI development, providing a unified API to simplify large model integration.
|
||||
|
||||
Our team uses Spring AI for Java projects. It unifies various large models' APIs into a single interface, allowing developers to easily switch between models without changing code.
|
||||
|
||||
|
||||
|
||||
Moreover, Spring AI offers many advanced features like conversation memory, RAG, and tool calling, enabling us to quickly build complex AI applications.
|
||||
|
||||
### LangChain4j
|
||||
|
||||
[LangChain4j](https://docs.langchain4j.dev/intro) is the Java version of LangChain, supporting declarative syntax and offering a great development experience, ideal for building complex agents.
|
||||
|
||||
I particularly like its chained call syntax, which is elegant to write. Its documentation is comprehensive, and the community is active, making it easy to find solutions on GitHub Issues.
|
||||
|
||||
When leading [Programming Navigation](https://codefather.cn/) members to develop an [AI No-Code Generation Platform Project](https://www.codefather.cn/course/1948291549923344386) comparable to major companies, we used LangChain4j + LangGraph4j, which made it easy to build complex AI workflows.
|
||||
|
||||
|
||||
|
||||
### Vercel AI Gateway
|
||||
|
||||
[Vercel AI Gateway](https://vercel.com/ai-gateway) is an AI model gateway service that allows unified access to hundreds of AI models.
|
||||
|
||||
It provides a unified API, so I don't need to learn each model's API documentation—I just use Vercel AI Gateway.
|
||||
|
||||
|
||||
|
||||
I use it to develop creative projects. For example, if I want to create an AI tool that lets users choose different models, Vercel AI Gateway makes it easy to implement this feature.
|
||||
|
||||
### Alibaba Cloud Bailian
|
||||
|
||||
[Alibaba Cloud Bailian](https://bailian.console.aliyun.com/) is an enterprise-level AI application development platform supporting RAG knowledge bases, process orchestration, and more.
|
||||
|
||||
Our team uses it to manage some knowledge bases. For example, we have many technical documents, and Bailian's RAG feature makes it easy for AI to answer questions based on these documents.
|
||||
|
||||
|
||||
|
||||
Moreover, Bailian offers visual process orchestration, allowing us to build complex AI workflows without writing code.
|
||||
|
||||
|
||||
|
||||
### OpenRouter
|
||||
|
||||
[OpenRouter](https://openrouter.ai/) is a unified AI model routing platform. It supports hundreds of models and allows easy switching without registering accounts on each platform.
|
||||
|
||||
Whenever a new model is released, I try it on OpenRouter to experience the latest capabilities.
|
||||
|
||||
|
||||
|
||||
For developers looking to experiment, OpenRouter is an excellent choice.
|
||||
|
||||
## 5. AI Browser Extensions
|
||||
|
||||
AI browser extensions allow me to use AI anytime within the browser, which is incredibly convenient.
|
||||
|
||||
### Monica
|
||||
|
||||
[Monica](https://monica.im/) is an all-in-one AI assistant in the browser, offering webpage summarization, video summarization, translation, writing, and image processing.
|
||||
|
||||
|
||||
|
||||
I use Monica daily. For example, when reading long articles, I use it to summarize the core content; when reading English documents, I use it for translation; and when browsing code repositories, I summarize the project introductions.
|
||||
|
||||
|
||||
|
||||
Monica's biggest advantage is its high integration, covering almost all common AI functions. Moreover, it supports multiple large models, allowing me to choose the most suitable one for each task. Highly recommended.
|
||||
|
||||
### AITDK
|
||||
|
||||
[AITDK](https://aitdk.com/) is an SEO and content creation tool. I use it for keyword analysis and competitor research.
|
||||
|
||||
Content creators know the importance of keyword research. AITDK helps me analyze which keywords have high search volume and low competition, so I can write articles or optimize my website accordingly.
|
||||
|
||||
Moreover, AITDK can analyze competitors' content strategies, showing what others are writing, helping me find unique angles.
|
||||
|
||||
|
||||
|
||||
## 6. AI Daily Conversations
|
||||
|
||||
Besides work, I also use AI to solve daily problems.
|
||||
|
||||
### Byte Doubao
|
||||
|
||||
Byte Doubao is a national-level AI application known for its speed and smooth interaction.
|
||||
|
||||
Whenever I encounter a problem, I often pull out my phone and ask Doubao. Its response speed is incredibly fast, and its Chinese comprehension is excellent.
|
||||
|
||||
### Tencent Yuanbao
|
||||
|
||||
Tencent Yuanbao's biggest advantage is its integration with WeChat, making it accessible anytime, anywhere.
|
||||
|
||||
When I see an article in WeChat and want an AI summary, I just @ Yuanbao. This seamless integration is fantastic—I can use AI within WeChat without switching apps.
|
||||
|
||||
|
||||
|
||||
## 7. AI Office Tools
|
||||
|
||||
When creating content, I also use AI office tools to improve efficiency.
|
||||
|
||||
### Baidu Wenxin Assistant
|
||||
|
||||
[Baidu Wenxin Assistant](https://image.baidu.com/) integrates many document writing and image processing tools. I often use it for image editing and watermark removal.
|
||||
|
||||
For example, if I take a photo and want to remove a watermark, Wenxin Assistant can do it in seconds. Or if I want to change a photo's background, Wenxin Assistant can handle that too.
|
||||
|
||||
Moreover, Wenxin Assistant offers many other features like PPT generation and document translation. It's a practical tool for content creators.
|
||||
|
||||
|
||||
|
||||
## 8. Yupi AI Navigation
|
||||
|
||||
The [AI Navigation Website](https://ai.codefather.cn/) developed by Yupi's team is also my most-visited AI website this year—it's comprehensive!
|
||||
|
||||
I regularly check for the latest AI trends. Moreover, Yupi AI Navigation categorizes many AI tools, making it easy to find what I need.
|
||||
|
||||
|
||||
|
||||
## 9. My Usage Strategy
|
||||
|
||||
After sharing so many tools, here's a quick summary:
|
||||
|
||||
- **For major projects**, I use Cursor + Claude Opus 4.5, as the code quality requirements are high, necessitating the strongest AI capabilities.
|
||||
|
||||
- **For small projects or quick idea validation**, I use Claude Code + Zhipu GLM, as they are fast and cost-effective.
|
||||
|
||||
- **When reading projects in VS Code**, I pair Cline + DeepSeek, as it's convenient.
|
||||
|
||||
- **For image generation**, I use Nano Banana, as it supports Chinese well and generates quickly.
|
||||
|
||||
- **For video generation**, I choose Jimeng AI (cost-effective) or Veo 3 (highest quality) based on quality needs.
|
||||
|
||||
- **For daily problem-solving**, I directly ask Doubao or Yuanbao, as they are quick and convenient.
|
||||
|
||||
- **When writing articles needing images**, I use Wenxin Assistant for image processing, as it's comprehensive.
|
||||
|
||||
Choosing the most suitable tool for each scenario ensures efficiency while controlling costs.
|
||||
|
||||
## Final Thoughts
|
||||
|
||||
These are my Top 20 most-used AI products this year, along with my practical insights.
|
||||
|
||||
**Tools will keep evolving, so flexibility in selection is key.**
|
||||
|
||||
Beyond tools, the core competence of Vibe Coding is more important: how to clearly express requirements, how to effectively communicate with AI, how to understand and optimize AI-generated code, and how to turn ideas into products.
|
||||
|
||||
Now that you've mastered the tool knowledge needed for Vibe Coding, it's time to use these tools to create real projects!
|
||||
|
||||
In the [Project Practice] section, I'll guide you through various projects, from personal tools to AI applications, from simple to complex, helping you truly grasp the essence of Vibe Coding.
|
||||
|
||||
## Recommended Resources
|
||||
|
||||
1) Yupi AI Navigation Website: [Comprehensive AI Resources, Latest AI News, Free AI Tutorials](https://ai.codefather.cn)
|
||||
|
||||
2) Programming Navigation Learning Circle: [Learning Paths, Programming Tutorials, Practical Projects, Job Hunting Guides, Q&A](https://www.codefather.cn)
|
||||
|
||||
3) Programmer Interview Cheatsheets: [Internship/Campus/Social Recruitment High-Frequency Topics, Enterprise Exam Analysis](https://www.mianshiya.com)
|
||||
|
||||
4) Programmer Resume Builder: [Professional Templates, Rich Examples, Direct to Interviews](https://www.laoyujianli.com)
|
||||
|
||||
5) 1-on-1 Mock Interviews: [Essential for Internship/Campus/Social Recruitment Interviews to Secure Offers](https://ai.mianshiya.com)
|
||||
Reference in New Issue
Block a user