L-Station Registration Guide: Check Account Age & Get Invite Link (No Essay Required)

The strongest AI community on Chinese Twitter, L-Station, is a paradise for freebie hunters. But many people still don't know how to register, so I built a small tool to help you check your account age and find the application entry point. You can apply for an invitation without writing an essay.

I posted about L-Station supporting clients two days ago, and it unexpectedly went viral. Many people in the comments asked how to register, so I wrote this article.

https://x.com/Lonely__MH/status/2107830806324375918

https://x.com/Lonely__MH/status/2063791578905485544

Actually, registration is simple now; you no longer need to write an essay: if your GitHub account is over 5 years old, you can apply for an invitation link and complete registration via that link.

I also made a small tool where you input your username to check if you meet the age requirement. If you do, it displays the official application entry point, which you can click to proceed.

The tool link will be in the comments 👇🏻

01 | Simple Three-Step Application

1. Check Age: Input your GitHub username or profile link into the tool to see if it's been over 5 years. No login required.
2. Get Invite: Once eligible, enter the official application portal and follow prompts to authorize GitHub.
3. Complete Registration: Go to GitHub to bind your email and receive the invite. Click the link and use that email to finish registration.

My account is already registered, so it shows "Registered". Remember these three rules:

• Links are valid for 30 minutes;
• Must register using the bound email;
• Each account can claim once every 3 days.

🚨 If the application entry doesn't open, try accessing with a domestic IP first. If you're using a proxy, add a direct connection rule for L-Station in Clash:

1DOMAIN-SUFFIX,linuxdo.org,DIRECT

⚠️ Don't click "GitHub Login" on the linux.do homepage—that's for logging in, not applying. Official applications are completely free; don't trust anyone selling invite codes.

The tool only predicts eligibility based on age; final qualification depends on official rules.

02 | Key Point: The Tool Was Built by AI

I didn't write a single line of code for the tool above.

I directly integrated Ant Group's Ling-3.1-flash into Pi, letting it read materials, write code, and test. Pi provides tools and execution environment, while Ling handles planning and progression.

Ling-3.1-flash is primarily designed for long-process development and complex Agent tasks. According to official docs, it has 560B total parameters, activating 25B per token, with 256K context currently available. The key feature is maintaining task state across multiple rounds of tool calls and adjusting based on results.

The prompt was just this paragraph:

1Help me create a mobile H5 page: Input GitHub username or profile link, query how long the account has been registered. If it reaches 5 years, automatically jump to Linux DO's invitation application entry. If not, tell the user how much time is left. Please read the following materials to verify rules, plan, implement, and run verification yourself, then deliver to me. Clearly mark any unconfirmed areas.
2
3GitHub User Query API: https://docs.github.com/en/rest/users/users#get-a-user
4Linux DO Registration Rules: https://linux.do/t/topic/2317378
5Official Application Entry: https://invite.linuxdo.org/

The actual version displays the application entry after querying, requiring users to click to proceed rather than auto-jumping.

Later, I added project profiles: select some public repos, let it read descriptions and READMEs, organize project intros and share cards. Let Ling write features during development, and let Ling read projects when using them.

Three aspects of the process impressed me most:

1. Reading Materials: Marks what it can't read

Some materials failed to fetch, so it switched to search and command-line requests. For those still unreadable, it lists "Unconfirmed" in the delivery, reminding users to rely on official pages for registration eligibility.

2. Verification: Opens browser to test itself

This surprised me most. After writing the function, it controls the browser itself, opens pages, operates, checks results, and debugs back—not just finishing code.

3. Debugging: Proposes causes, then finds ways to verify

When analyzing repos, it reported "Invalid model output format." It checked the calling code, found max_tokens set to 3000, suspected reasoning consumed output quota causing JSON truncation, and prepared to compare returns under different quotas.

During testing, it hit API rate limits and had to wait for recovery. Whether errors were due to truncation needs re-testing confirmation.

I originally just wanted a small tool, but didn't expect it to handle debugging and verification too. This part impressed me more than the final page appearance.

03 | Integration & Free Trial

Ling-3.1-flash currently offers a two-week free trial, giving 500k tokens daily. L-Station friends should like this 😄

• Web Version 👉🏻 Ling Studio
• API 👉🏻 Ant Digital MaaS
• OpenRouter 👉🏻 Ling 3.1 Flash, Model ID: inclusionai/ling-3.1-flash

🚨 Keys and addresses must match: Use MaaS keys with maas-api.antdigital.com. Filling them into other Ling addresses reports "Invalid Authorization Token." I got burned here 😄, damn it.

To use it in Pi like me, install first:

1npm install -g --ignore-scripts @earendil-works/pi-coding-agent

Add Ling to ~/.pi/agent/models.json. Merge fields if existing configs exist.

1{
2  "providers": {
3    "ant-ling": {
4      "baseUrl": "https://maas-api.antdigital.com/v1",
5      "api": "openai-completions",
6      "models": [
7        {
8          "id": "ling-3.1-flash",
9          "name": "Ling 3.1 Flash",
10          "reasoning": true,
11          "input": ["text"],
12          "contextWindow": 262144,
13          "maxTokens": 32768,
14          "thinkingLevelMap": {
15            "off": null,
16            "minimal": null,
17            "low": null,
18            "medium": null,
19            "high": "high",
20            "xhigh": "xhigh"
21          }
22        }
23      ]
24    }
25  }
26}

Create a Key in MaaS, fill it into ~/.pi/agent/auth.json:

1{
2  "ant-ling": { "type": "api_key", "key": "Your API Key" }
3}

Set as default in ~/.pi/agent/settings.json:

1{
2  "defaultProvider": "ant-ling",
3  "defaultModel": "ling-3.1-flash"
4}

Type pi in project directory, ask a question, get normal reply = connected ✌🏻. Config docs at Pi Official Docs.

💡 Keys and API addresses must correspond: Pair MaaS keys with maas-api.antdigital.com, don't mix with other platforms' keys.

🔥 Tool Address: https://ldo.aigce.com

📚 Past Articles

1. 🔥 Infinite Canvas, More Than Image Gen
2. What's it like having all terminal Agents on one phone?
3. Complete Guide to Local LLM Deployment
4. Hermes Agent Practical Guide: From X Anxiety to Auto-Sinking
5. Programmer Hair Loss Prevention Guide
6. Hermes Connects to iMessage
7. Hermes Connects to X Premium
8. Hermes Agent Complete Guide
9. Hermes Agent Intro: Auxiliary Models
10. Hermes Agent Intro Guide
11. Hermes Agent Advanced Guide
12. Hermes Agent Incomplete Guide
13. Nigeria Region Claude Pro Subscription Complete Guide
14. Nigeria Region Apple ID Registration Tutorial
15. Turkey Region Half-Price ChatGPT Plus Subscription
16. US Region Apple ID Registration
17. Claude/ChatGPT/Gemini Alipay Subscription
18. Mac Local LLM Deployment Complete Tutorial
19. IP Quality Check
20. Why Doubao Doesn't Recommend Your Brand
21. How to Explain to Your Grandma That What Doubao Says Isn't True

If this tutorial helped you, please Follow + Like + Retweet 👏🏻

Follow @Lonely__MH for beginner-friendly tutorials and AI tool insights.