[GitHub Global] Translate Vibe Coding 零基础教程/20 项目实战/鱼皮的原创项目/Kimi K2 - AI 文档阅读助手项目实战.md to en
This commit is contained in:
yupi-translate-app[bot]
committed by
GitHub
GitHub
parent
78ec312463
commit
2b7bd89414
@@ -0,0 +1,133 @@
|
||||
# Kimi K2 - AI Document Reading Assistant Project Practice
|
||||
|
||||
This project is an AI document reading assistant website that helps you quickly understand various complex documents (papers, technical documents, PDFs, etc.) and also helps you manage your documents.
|
||||
|
||||
The project includes a complete frontend and backend, developed entirely through conversations with AI, without writing a single line of code. It is suitable for those who want to quickly practice the complete Vibe Coding development process and learn how to develop practical tools using AI.
|
||||
|
||||
---
|
||||
|
||||
Hello everyone, I am programmer Yupi. The school season has arrived, and I believe many friends are starting to collect and read papers. Like me, when learning new technical knowledge, I also read documents, and I deeply understand the pain of reading documents. Even though each word seems understandable when separated, when put together, it becomes incomprehensible.
|
||||
|
||||
|
||||
|
||||
To help everyone avoid the torment of reading documents, I developed an AI document assistant website using AI, which can help you quickly understand various complex documents and also manage your documents.
|
||||
|
||||
|
||||
|
||||
The website is completely free, and the code is fully open source!
|
||||
|
||||
Open-source repository: [github.com/liyupi/literature-assistant](https://github.com/liyupi/literature-assistant)
|
||||
|
||||
|
||||
|
||||
Below, I will first teach you how to use the website, then share the process of making this website, and also introduce the method of using Claude Code in China.
|
||||
|
||||
⭐️ Recommended to watch the video version, learn in 2 minutes: [bilibili.com/video/BV1MnpVzdETW](https://www.bilibili.com/video/BV1MnpVzdETW/)
|
||||
|
||||
## How to Use?
|
||||
|
||||
First, download the open-source code to your computer, then directly run the quick start script I provided, and you can see the effect by opening the webpage.
|
||||
|
||||
💡 Make sure your computer has Node.js and Java environments installed. You can refer to the README.md document for installation instructions.
|
||||
|
||||
|
||||
|
||||
When you want to read a document, click the "Single Import" button, upload the document file, and then you need to fill in the Kimi AI API Key.
|
||||
|
||||
|
||||
|
||||
I chose Kimi because they just released the new version of the K2 model, which performs well in programming, reasoning, and document understanding;
|
||||
|
||||
Moreover, it supports a context of 256K, so it can handle papers with hundreds of thousands of words.
|
||||
|
||||
|
||||
|
||||
In benchmark tests like SWE-bench Verified, which focus on real software engineering tasks, the new Kimi K2 model also performs well:
|
||||
|
||||
|
||||
|
||||
Just log in to the [Kimi Developer Console](https://platform.moonshot.cn/), then go to API Key management to get a key for calling the large model.
|
||||
|
||||
|
||||
|
||||
Although new users have free credits, do not leak your key!
|
||||
|
||||
|
||||
|
||||
After filling in the API Key, you can generate the document reading guide, and the generation speed is very fast.
|
||||
|
||||
|
||||
|
||||
The effect generated by AI is quite good, with both text and images, helping you understand complex documents faster.
|
||||
|
||||
|
||||
|
||||
You can also batch import multiple documents and simultaneously call AI to generate reading guides, improving efficiency.
|
||||
|
||||
|
||||
|
||||
In addition, you can use this website as your smart document collection folder, where you can categorize and retrieve imported documents, download original files, and view document reading guides at any time. **Don't let your collected documents get lost again~**
|
||||
|
||||
|
||||
|
||||
## How to Implement?
|
||||
|
||||
In the past, such a website might take several days to build. But now, AI programming technology is mature. I chose Claude Code AI development tools and easily completed it in one day without writing a single line of code.
|
||||
|
||||
First, enter a command in the terminal to install Claude Code:
|
||||
|
||||
```bash
|
||||
npm install -g @anthropic-ai/claude-code
|
||||
```
|
||||
|
||||
Then execute the `claude` command, and you can ask it questions~
|
||||
|
||||
Result: an error occurred!
|
||||
|
||||
|
||||
|
||||
**Damn it, this thing doesn't support use in China yet!**
|
||||
|
||||
But it's okay, we can switch to Kimi. Enter a command in the terminal to configure an environment variable (note the difference between operating systems):
|
||||
|
||||
```bash
|
||||
# Linux/macOS start the high-speed version of the kimi-k2-turbo-preview model
|
||||
export ANTHROPIC_BASE_URL=https://api.moonshot.cn/anthropic
|
||||
export ANTHROPIC_AUTH_TOKEN=<Your API Key>
|
||||
export ANTHROPIC_MODEL=kimi-k2-turbo-preview
|
||||
export ANTHROPIC_SMALL_FAST_MODEL=kimi-k2-turbo-preview
|
||||
|
||||
# Windows Powershell start the high-speed version of the kimi-k2-turbo-preview model
|
||||
$env:ANTHROPIC_BASE_URL="https://api.moonshot.cn/anthropic";
|
||||
$env:ANTHROPIC_AUTH_TOKEN=<Your API Key>
|
||||
$env:ANTHROPIC_MODEL="kimi-k2-turbo-preview"
|
||||
$env:ANTHROPIC_SMALL_FAST_MODEL="kimi-k2-turbo-preview"
|
||||
```
|
||||
|
||||
Then you can happily use Claude Code to generate code~
|
||||
|
||||
|
||||
|
||||
For a website that includes a complete frontend and backend, it's difficult to generate satisfactory results with just one prompt. Therefore, we need to **decompose the work steps** like real enterprise development. First, the backend, then the frontend, and finally, connect and debug the frontend and backend. It's best to develop features one by one and adjust promptly if problems arise.
|
||||
|
||||
Here are some reference prompts:
|
||||
|
||||
|
||||
|
||||
---
|
||||
|
||||
That's all for this sharing. I hope this tool is helpful to everyone, and don't forget to support Yupi with likes, shares, and comments. Thank you all~
|
||||
|
||||
|
||||
|
||||
## Recommended Resources
|
||||
|
||||
1) Yupi AI Navigation Website: [AI Resource Collection, Latest AI News, Free AI Tutorials](https://ai.codefather.cn)
|
||||
|
||||
2) Programming Navigation Learning Circle: [Learning Paths, Programming Tutorials, Practical Projects, Job Hunting Guide, Q&A](https://www.codefather.cn)
|
||||
|
||||
3) Programmer Interview Cheatsheet: [Internship/Campus Recruitment/Social Recruitment High-Frequency Exam Points, Enterprise Real Questions Analysis](https://www.mianshiya.com)
|
||||
|
||||
4) Programmer Resume Writing Tool: [Professional Templates, Rich Examples, Direct to Interview](https://www.laoyujianli.com)
|
||||
|
||||
5) 1-on-1 Mock Interview: [Essential for Internship/Campus Recruitment/Social Recruitment Interviews to Get Offers](https://ai.mianshiya.com)
|
||||
Reference in New Issue
Block a user