[GitHub Global] Translate Vibe Coding 零基础教程/10 编程工具/工具实战/OpenCode：开源免费的 AI 命令行工具实测.md to en
This commit is contained in:
yupi-translate-app[bot]
committed by
GitHub
GitHub
parent
ff2ab4d95c
commit
bca2730634
@@ -0,0 +1,392 @@
|
||||
# OpenCode: Hands-on Review of an Open-Source Free AI Command Line Tool
|
||||
|
||||
Hello everyone, I'm programmer Yupi.
|
||||
|
||||
Claude Code has long been recognized as the undisputed Top 1 AI programming command line tool, enjoying an almost god-like status in the AI and programmer communities.
|
||||
|
||||
|
||||
|
||||
But this damn thing isn't very friendly to Chinese users...
|
||||
|
||||
First, to use Claude Code, you must have special network access + an official account, otherwise you'll just see red errors.
|
||||
|
||||
|
||||
|
||||
Moreover, in September 2025, Anthropic suddenly announced **a complete ban on Chinese-controlled enterprises using Claude services** for some inexplicable reason. This ban not only includes mainland Chinese companies but also overseas companies with over 50% Chinese ownership!
|
||||
|
||||
Anthropic even specifically named China, calling us an **adversarial nation**!
|
||||
|
||||
|
||||
|
||||
The world has suffered under Claude Code's dominance for too long!
|
||||
|
||||
But recently, many programmer friends around me have started switching from Claude Code to another tool - the suddenly popular open-source project OpenCode.
|
||||
|
||||
|
||||
|
||||
In just half a year, this project has skyrocketed to 52k stars on GitHub!
|
||||
|
||||
To put this in perspective - that's more stars than all my dozens of open-source projects on GitHub combined! So jealous...
|
||||
|
||||
|
||||
|
||||
What exactly is OpenCode? Why is it so popular?
|
||||
|
||||
|
||||
|
||||
## What is OpenCode?
|
||||
|
||||
[OpenCode](https://opencode.ai/) is a 100% open-source AI programming command line tool that can be used in **terminals, IDEs, and even desktop applications**.
|
||||
|
||||
|
||||
|
||||
You might ask: How is this different from Claude Code?
|
||||
|
||||
Why not try it and see?
|
||||
|
||||
Next, I'll walk you through a hands-on demo - from installation and configuration to actual coding, all in one go~
|
||||
|
||||
|
||||
|
||||
## Getting Started with OpenCode from Scratch
|
||||
|
||||
### 1. Installing OpenCode
|
||||
|
||||
Go directly to the OpenCode official website and copy this command:
|
||||
|
||||
|
||||
|
||||
Here's the command:
|
||||
|
||||
```bash
|
||||
curl -fsSL https://opencode.ai/install | bash
|
||||
```
|
||||
|
||||
Execute it in your terminal to complete the installation.
|
||||
|
||||
After installation, enter `opencode` to launch the program, and you're ready to use it happily~
|
||||
|
||||
|
||||
|
||||
Let's start with the classic Hello World - the AI successfully responded.
|
||||
|
||||
|
||||
|
||||
Congratulations, you've now mastered 70% of OpenCode.
|
||||
|
||||
### 2. Selecting Modes and Models
|
||||
|
||||
OpenCode supports 2 modes. The default is Build mode for application development and code generation.
|
||||
|
||||
Press Tab to switch to Plan mode for generating execution plans.
|
||||
|
||||
|
||||
|
||||
Press `Ctrl + p` to open the command palette with dozens of built-in commands. Let's try switching the LLM first:
|
||||
|
||||
|
||||
|
||||
By default, it offers 4 free models:
|
||||
|
||||
|
||||
|
||||
Wow, even the latest GLM-4.7 from Zhipu is free? Did I waste money on my Coding Plan subscription?
|
||||
|
||||
|
||||
|
||||
Besides free models, OpenCode supports a huge number of AI models for you to choose freely:
|
||||
|
||||
|
||||
|
||||
After selecting a model, just configure your API Key:
|
||||
|
||||
|
||||
|
||||
If you previously had a **Claude Pro/Max subscription account**, you can log in directly and seamlessly migrate from Claude Code.
|
||||
|
||||
|
||||
|
||||
### 3. Quick Commands
|
||||
|
||||
OpenCode supports slash commands. Type `/` to see many operations like viewing model lists, checking Agents, managing MCP, switching themes, etc.:
|
||||
|
||||
|
||||
|
||||
It supports dozens of different themes, all quite aesthetically pleasing. This shows OpenCode really cares about user experience:
|
||||
|
||||
|
||||
|
||||
Type `@` to quickly associate directory files and add context for the AI:
|
||||
|
||||
|
||||
|
||||
### 4. Interactive Experience
|
||||
|
||||
Compared to Claude Code, OpenCode really maximizes the command line interactive experience. I'd say it's more like a desktop app disguised as a command line tool.
|
||||
|
||||
You can click any message to pop up an action box where you can recall messages and AI replies, copy them, or start a new dialog based on the current conversation.
|
||||
|
||||
|
||||
|
||||
You can scroll up/down to switch menus and directly click to proceed to the next step.
|
||||
|
||||
Press `Ctrl + p` to open the command palette and enable the sidebar:
|
||||
|
||||
|
||||
|
||||
Then the interface becomes like this. You call this a command line?
|
||||
|
||||
|
||||
|
||||
### 5. LSP Support
|
||||
|
||||
If you're observant, you've noticed the `LSP` in the right sidebar. What's this? Some perverted thing?
|
||||
|
||||
LSP (Language Server Protocol) is a communication protocol developed by Microsoft to enable communication between code editors and language servers.
|
||||
|
||||
Simply put, **LSP is the technology that helps editors understand code.**
|
||||
|
||||
For example, when you type `console.` in VS Code, it automatically suggests `log`. You can jump to definitions by clicking function names, and incorrect code gets red underlines. These editor features all rely on LSP.
|
||||
|
||||
OpenCode's LSP support means the AI truly understands your code structure rather than treating code as plain text for blind guessing, making modifications more precise.
|
||||
|
||||
For instance, when I ask the AI to introduce the most valuable code in my AI quiz platform project, LSP comes into play. It helps the AI quickly locate where a piece of code is called and what variables it references, instead of having the AI dumbly search through text globally.
|
||||
|
||||
|
||||
|
||||
### 6. Returning to Previous Sessions
|
||||
|
||||
If you accidentally close OpenCode, don't worry. Open the command palette and select "Switch session":
|
||||
|
||||
|
||||
|
||||
You can return to your previous chat:
|
||||
|
||||
|
||||
|
||||
## Desktop Version of OpenCode
|
||||
|
||||
Even with all these user experience improvements in OpenCode, I suspect most of you still dislike the black terminal box.
|
||||
|
||||
No problem - OpenCode also offers a desktop app version! Supporting macOS, Windows, and Linux across all platforms. They're really going all out to compete with Claude Code...
|
||||
|
||||
> Link: https://opencode.ai/download
|
||||
|
||||
|
||||
|
||||
But when I installed and opened it with great enthusiasm, it errored out!
|
||||
|
||||
|
||||
|
||||
After some troubleshooting, I found it was because I had a proxy enabled. After disabling it, it ran normally.
|
||||
|
||||
|
||||
|
||||
But having gotten used to Cursor, this interaction experience feels a bit perfunctory. I don't recommend using it.
|
||||
|
||||
## OpenCode's Extensibility
|
||||
|
||||
So far, I think OpenCode completely crushes Claude Code in frontend user experience. Moreover, OpenCode is fully compatible with Claude Code's Skills system!
|
||||
|
||||
Skills are capability extension packages prepared for the AI. You can think of them as onboarding documents for new colleagues, containing task execution methods, tool usage instructions, template materials, etc.
|
||||
|
||||
For example, you can create a `Company Code Style Skill` documenting code styles, naming conventions, comment requirements, etc. Afterward, Claude Code will automatically follow these standards when generating code without needing repeated explanations.
|
||||
|
||||
According to official documentation, OpenCode automatically searches for Skills in these locations:
|
||||
|
||||
- `.opencode/skill/<name>/SKILL.md` (project directory)
|
||||
- `~/.config/opencode/skill/<name>/SKILL.md` (user directory)
|
||||
- `.claude/skills/<name>/SKILL.md` (Claude Code compatible)
|
||||
- `~/.claude/skills/<name>/SKILL.md` (Claude Code compatible)
|
||||
|
||||
This means if you've previously created custom Skills for Claude Code, you can use them directly with OpenCode! Another seamless migration.
|
||||
|
||||
## Oh My OpenCode Supercharged Plugin
|
||||
|
||||
If you think OpenCode isn't powerful enough, try the open-source OpenCode enhancement plugin `Oh My OpenCode`, already with 10k stars.
|
||||
|
||||
> Project address: https://github.com/code-yeongyu/oh-my-opencode
|
||||
|
||||
|
||||
|
||||
How amazing is this plugin? Check out user reviews:
|
||||
|
||||
> "It made me cancel my Cursor subscription."
|
||||
>
|
||||
> "Knocked out 8000 eslint warnings with Oh My Opencode, just in a day"
|
||||
|
||||
The core feature of Oh My OpenCode is introducing an agent orchestration system called **Sisyphus**.
|
||||
|
||||
I looked it up:
|
||||
|
||||
> Sisyphus is a king in Greek mythology punished by the gods for deceiving them and challenging authority. His punishment was endlessly pushing a boulder up a mountain, only for it to roll back down upon reaching the top, symbolizing futile, never-ending tasks and representing a spirit of rebellion against absurd fate.
|
||||
|
||||
This system can:
|
||||
|
||||
1. Schedule multiple AI models in parallel: e.g., have GPT debug while Gemini writes frontend code
|
||||
2. Automatic task management: Won't stop until tasks are completed, persevering like Sisyphus pushing the boulder
|
||||
3. Intelligent code review: Automatically detects and cleans redundant AI-generated comments
|
||||
4. Deep LSP integration: Provides IDE-level features like renaming and jumping to definitions
|
||||
|
||||
In short, Sisyphus is an AI supervisor that can command multiple AI models simultaneously while ensuring they complete tasks.
|
||||
|
||||
|
||||
|
||||
Although the official docs say installation can be done with one command, I recommend first installing bun, then using npx to install, otherwise errors may occur.
|
||||
|
||||
```bash
|
||||
npm install bun -g
|
||||
npx oh-my-opencode install
|
||||
```
|
||||
|
||||
During installation, it may ask if you have subscriptions to certain models. I just kept selecting "No":
|
||||
|
||||
|
||||
|
||||
After installation completes, enter OpenCode again. Then just include the `ultrawork` (or `ulw`) cheat code in your prompts to activate all enhanced features - automatic scheduling of multiple AI models working simultaneously, deep codebase exploration, and relentless execution.
|
||||
|
||||
Let's test this out and see if OpenCode can really kick Claude Code to the curb when it comes to project capabilities?
|
||||
|
||||
## Hands-on Project - Building an AI Health Assistant with OpenCode
|
||||
|
||||
Recently, Ant Group's `Ant Afu` AI health assistant went viral, with ads featuring host He Jiong appearing everywhere from subway stations to office building TVs.
|
||||
|
||||
|
||||
|
||||
Although I haven't used it yet, I heard it can provide AI preliminary diagnoses by scanning skin or reports, along with intelligent answers to medical questions and treatment suggestions.
|
||||
|
||||
Let's build a similar health assistant website!
|
||||
|
||||
Before there was Ant Afu, now there's Yupi Akun.
|
||||
|
||||
|
||||
|
||||
First, let's analyze: we're building a full-stack project including frontend + backend, with the backend needing to call AI models to generate content.
|
||||
|
||||
Here I chose to use **Vercel AI Gateway** to implement AI capabilities - a simple and easy-to-use AI gateway.
|
||||
|
||||
|
||||
|
||||
What's an AI gateway?
|
||||
|
||||
Simply put, it's like a train station ticket gate. Your application's requests first pass through the gateway, which handles authentication, rate limiting, monitoring, and other complex operations before forwarding requests to AI models.
|
||||
|
||||
|
||||
|
||||
Moreover, Vercel AI Gateway supports integration with over 500 large models and has free quotas, making it perfect for learning and small projects.
|
||||
|
||||
> Link: https://vercel.com/ai-gateway
|
||||
|
||||
1) First, you need to register/login to Vercel, then create an API Key in the console. Be careful not to leak it:
|
||||
|
||||
|
||||
|
||||
2) Launch OpenCode, switch the model to the highly capable and free GLM-4.7, then input this prompt:
|
||||
|
||||
```markdown
|
||||
You are a professional programmer. Please help me develop the "Daily Health Assistant" website where users can chat with AI to record and manage daily health status.
|
||||
|
||||
## Development Requirements
|
||||
|
||||
1. Must include complete frontend and backend, backend using Node.js
|
||||
2. Implement AI capabilities using Vercel's AI Gateway, first check official docs for usage: https://vercel.com/docs/ai-gateway/getting-started
|
||||
3. Focus on completing core functionality to ensure the project runs properly
|
||||
4. Overall website interface should adopt a fresh green health style, responsive across devices
|
||||
5. AI should proactively ask about user health status like sleep, exercise, diet, etc.
|
||||
```
|
||||
|
||||
After sending, OpenCode will automatically use web scraping tools to read Vercel AI Gateway's official docs and learn the latest usage:
|
||||
|
||||
|
||||
|
||||
In about 5 minutes, the AI completed all code generation and automatically installed dependencies.
|
||||
|
||||
|
||||
|
||||
3) I directly provided the Vercel API Key I obtained earlier to the AI to help launch the project:
|
||||
|
||||
|
||||
|
||||
4) After successfully launching the project, open a browser to `localhost:3000` to test it.
|
||||
|
||||
But it errored! Couldn't call the AI.
|
||||
|
||||
|
||||
|
||||
Perhaps the AI misunderstood the Vercel AI Gateway documentation, leading to incorrect AI calling code. So I input the docs to the AI again for another attempt:
|
||||
|
||||
|
||||
|
||||
It errored again. Even though I provided the API Key, the system still reported "Missing API Key".
|
||||
|
||||
So I called the AI again, telling it "I already gave you this key earlier".
|
||||
|
||||
|
||||
|
||||
After about 5 rounds of errors and fixes, it still wouldn't work properly! I'm exhausted...
|
||||
|
||||
|
||||
|
||||
Then I had a mischievous idea: Since we're comparing to Claude Code, why not try using Claude Code to fix this problem OpenCode couldn't solve?
|
||||
|
||||
|
||||
|
||||
Let's try! Input prompt:
|
||||
|
||||
```markdown
|
||||
Currently the project's backend AI functionality isn't working
|
||||
Please refer to https://vercel.com/docs/ai-gateway/getting-started documentation
|
||||
Help fix the backend to ensure the project runs properly
|
||||
```
|
||||
|
||||
|
||||
|
||||
Claude Code successfully fixed the issue, and it finally worked normally:
|
||||
|
||||
|
||||
|
||||
💡 Note: If you encounter AI call network timeout issues, have the AI change the calling baseURL to https://ai-gateway.vercel.sh/v1
|
||||
|
||||
Previously, similar tasks using Claude Code/Cursor + GLM took under 10 minutes to complete. This time it took about 20 minutes with back-and-forth before working properly.
|
||||
|
||||
This makes me doubt OpenCode's capabilities. And it feels like the GLM model became dumber in OpenCode - or is that just my imagination...
|
||||
|
||||
No way - everyone's raving about OpenCode. I must be using it wrong!
|
||||
|
||||
|
||||
|
||||
### Ultrawork Mode
|
||||
|
||||
Remember the `ultrawork` (or `ulw`) cheat code mentioned earlier? Let's try it!
|
||||
|
||||
|
||||
|
||||
Entering battle mode:
|
||||
|
||||
|
||||
|
||||
You can view sub-agent operation details. First press `Ctrl + x`, then arrow keys to check different agents.
|
||||
|
||||
When background tasks complete, there's a notification. Here the "Research Vercel AI SDK Conversation Mode" task is done.
|
||||
|
||||
|
||||
|
||||
But guess what? After waiting nearly 10 minutes, the task still wasn't finished...
|
||||
|
||||
Looking at this task list - does it really need to be this complex? It even pulled in databases?
|
||||
|
||||
|
||||
|
||||
I've lost patience waiting. Just end it!
|
||||
|
||||
Apparently, this not-too-complex work doesn't leverage multi-agent advantages well. It's like needing to print one sheet of paper but mobilizing the entire company - some researching paper types, some studying printer status, some exploring optimal printing postures.
|
||||
|
||||
## Final Thoughts
|
||||
|
||||
After these simple tests, I'm keeping OpenCode under observation for now.
|
||||
|
||||
The frontend is indeed excellent, but backend capabilities seem behind Claude Code.
|
||||
|
||||
If I just want frontend convenience, why not use Cursor?
|
||||
|
||||
![](https://
|
||||
Reference in New Issue
Block a user