Build Your Personal Knowledge Base with WorkBuddy + ima: A Step-by-Step Tutorial

Many people's first reaction when mentioning a personal knowledge base is Obsidian.

It is certainly great: local Markdown, backlinks, plugins, directories, templates, and graphs—it is indeed powerful once you master it.

But the problem is, most ordinary people aren't stuck because their "knowledge base isn't advanced enough."

They are stuck because: their materials are too scattered.

A bunch of articles in WeChat favorites, documents in Tencent Docs, PDFs in computer folders, meeting minutes everywhere, videos watched without transcripts, and inspirations written in memos—eventually, no one can find anything.

At this point, asking a beginner to learn Obsidian, Markdown, templates, backlinks, and plugins is actually quite discouraging.

So, if your primary materials are within the Chinese office environment, I suggest looking at a combination more suitable for domestic users:

✅ WorkBuddy + ima

• ima is responsible for putting materials into the knowledge base.
• WorkBuddy is responsible for reading, organizing, categorizing, summarizing, and generating reviews.
• Humans are responsible for the final judgment: which content is worth keeping and which is just temporary material.

This is my understanding of building a personal knowledge base from 0 to 1.

Don't worry about building a complex system first; just let someone help you process the materials scattered in WeChat, documents, PDFs, meetings, and videos for the first time.

1. Why Not Start with Obsidian?

Obsidian's advantages are obvious: it's suitable for long-term accumulation, building your own structure, and weaving knowledge into a network.

But it has a practical problem: it is more like a "powerful warehouse."

But the warehouse itself won't help you organize the shelves.

Many people's Obsidian eventually becomes a second trash can instead of a second brain. Web clippings, PDFs, daily notes, and meeting minutes are all dumped in. It feels productive at first, but after a while, you don't even want to open it.

The reason is simple: no one does the second step of processing after the materials come in.

• Should a temporary thought be kept?
• Which project is an article related to?
• Which pages of a PDF are useful?
• What conclusions were actually drawn from a meeting?
• Can a video's content be turned into a topic?

If these questions aren't solved, the larger the knowledge base, the heavier the burden.

So, for ordinary people building a personal knowledge base, don't pursue "advanced" features in the first step.

Pursue one thing first: let someone help you with the preliminary processing after materials enter.

This is the value of WorkBuddy + ima.

2. How Does This Combination Divide Labor?

I break it down into three roles:

This division is important. Many people want AI to take over the entire process—automatic reading, categorizing, summarizing, archiving, and generating formal notes. It sounds great but is actually dangerous ⚠️.

Because long-term knowledge is not formatted text; it requires judgment.

• Which materials are worth keeping?
• Which viewpoints only look correct?
• Which content will actually be used in the future?
• Which materials felt useful at the time but will never be opened again?

These cannot be fully handed over to AI.

A more stable way is:

WorkBuddy handles pre-processing, ima handles accumulation, and humans handle confirmation.

Don't mess up this order. If you do, the knowledge base will become a pile of AI-generated beautiful nonsense.

3. Starting from 0: Connect in 5 Minutes

Step 1: Install and Log in to ima

1. Open https://ima.qq.com/ in your browser.
2. Download the Mac or Windows client and log in via WeChat scan.
3. Mobile: Search for the "IMA Knowledge Base" mini-program in WeChat and log in with the same account.

Important Reminder: ima and WorkBuddy must use the same WeChat account, or the knowledge base won't be readable.

Step 2: Connect ima in WorkBuddy

1. Open the WorkBuddy client.
2. Click "Library" on the left sidebar and select "ima Knowledge Base."
3. Click "Go to Authorize."
4. Scan the WeChat QR code to log in to the ima knowledge base.
5. Click "Confirm Authorization."

Once authorized, WorkBuddy will gain the following permissions:

• View your knowledge base list in WorkBuddy.
• View or search knowledge base materials.
• Add knowledge base files to tasks or save files to the ima knowledge base.

Verify Connection:

Type "Help me list all knowledge base names in ima" in the WorkBuddy dialog. If it returns your list, the connection is successful.

Step 3: Understand the Two Reference Methods

After connecting, there are two ways to reference ima content in tasks:

Method A — Add from Library: Select the file in "Library → ima Knowledge Base" and click "Add to Task."

Method B — Add in Dialog: Click the "Upload File" icon above the input box in a new task, open the ima knowledge base, select specific files or the whole base, and click "OK."

Step 4: Understand the Save-Back Function

After WorkBuddy processes materials, the output can be saved back to ima:

1. Find the file in the right-side output area.
2. Click "Upload to Cloud."
3. Select "ima Knowledge Base" for storage.
4. Supports Personal / Shared / Subscribed categories.

This creates a complete loop: Read → Process → Save Back.

4. Beginner Version: Only Create Three Knowledge Bases

Many tutorials make things too complex from the start. Resource bases, project bases, domain bases, card bases, output bases, archive bases...

It sounds professional, but for beginners, more categories mean a higher chance of giving up.

For the first version, only build three:

01-Resource Base

Used for external materials: articles, PDFs, reports, web pages, courses, video transcripts, interview notes, and meeting recordings.

02-Project Base

Used for things you are currently pushing forward: a social media topic, a client project, a product idea, a course plan, or a business proposal.

03-Output Base

Used for your own results: article drafts, topic pools, reviews, summaries, viewpoint cards, and methodology accumulations.

Why only three?

Because the most important thing for a beginner is not fine categorization, but developing a habit: don't leave materials scattered. Get them into a place where AI can read and process them.

Once this step works, you've already surpassed most people's "favorites folder knowledge base."

How to create a knowledge base in ima:

Open ima client → Left side "Knowledge Base" → Click "+" → Enter name → Done.

5. Don't Start with Too Much Data

The biggest mistake is trying to dump everything on day one. Hundreds of WeChat favorites, browser bookmarks, and PDFs. It looks impressive, but it's just moving a trash pile into a new house.

The first batch should be small. I suggest fewer than 20 items:

Why so few? Because the first stage isn't about "building a big knowledge base," but verifying three things:

1. Can ima store these materials well?
2. Can WorkBuddy read and organize them well?
3. Are you willing to continue using it after seeing the AI results?

If these aren't verified, more data just means more chaos. Test with a small batch of real data first.

How to Add Materials to ima:

• WeChat Articles: Copy link → Open ima client/mini-program → Enter knowledge base → Paste link.

• WeChat Files: Upload "WeChat Files" in the knowledge base → Select contact → Select file.
• Local Files: Open ima client → Enter knowledge base → Upload file.
• Photos/Screenshots: Open IMA mini-program → "+" → Take photo → Auto OCR.

Note: Sending files directly in the dialog is only for temporary files; you must upload them inside a "Knowledge Base" to truly store them.

6. Run Three Minimal Tasks After Connecting

Don't rush into automation. Run these three tasks first to establish the prototype.

Task 1: Organize Resource Base

Tell WorkBuddy:

1Please read "01-Resource Base" in the ima knowledge base.
2
3Help me organize recently uploaded materials.
4
5Group the output by theme.
6
7Each item should include: Title, one-sentence summary, core viewpoints, suitable theme, and whether it's worth intensive reading.
8
9Do not generate a final article.
10Do not delete, move, or overwrite any materials for me.

This turns a pile of data into a list you can scan quickly. You'll know what's worth reading and what's just noise.

Task 2: Generate Project Material Pack

1Please read "02-Project Base" in the ima knowledge base.
2
3Organize a material pack around "My Current Project."
4
5Please output: Project background, existing materials, key questions, available assets, missing information, and next steps.
6
7Do not make final decisions for me.
8Do not generate a formal proposal.

This pulls scattered project info into one view so you can see what's missing.

Task 3: Generate Content Topics

1Please read "01-Resource Base" and "03-Output Base" in the ima knowledge base.
2
3Help me generate 10 topics suitable for WeChat or X.
4
5Each topic includes: Title, core viewpoint, citable materials, suitable angle, and risk points.
6
7Do not write the full article directly.

7. The Correct Workflow: Data Doesn't Become Articles Directly

Avoid the trap of: Data in → AI reads → Generate article → Copy/Paste. This leads to generic content.

A more stable workflow:

1Data enters ima
2    ↓
3WorkBuddy generates review
4    ↓
5Human filters valuable content
6    ↓
7Generate topics
8    ↓
9Generate outline
10    ↓
11Finally, write the text

8. The Core is "Reuse," Not "Storage"

If you never look at it again, it's just a favorites folder. I suggest a weekly review:

1Please read new materials added to the ima knowledge base this week.
2
3Generate a weekly review.
4
5Include: New content, recurring themes, materials worth intensive reading, topics that can be developed into articles, information needed, and action suggestions for next week.
6
7Do not delete, move, or overwrite any materials.

9. Remote Triggers: Process Tasks Away from Your Desk

WorkBuddy can be triggered via WeChat, DingTalk, or Feishu. Use this for low-risk tasks like searching for info or summarizing a meeting you just finished.

10. What Materials to Put in ima?

• Articles: WeChat, web pages, industry analysis.
• PDFs: Reports, white papers, courseware.
• Meeting Minutes: Transcripts and notes.
• Video Transcripts: Courses, interviews, podcasts.
• Your Inspirations: Your own judgments are the most valuable.

11. What NOT to Put in Directly

• Un-anonymized client data.
• Sensitive info (contracts, IDs, bank cards).
• Company secrets.
• High-risk judgment materials (medical, legal, investment).

12. Keep Naming Simple

13. 5 Prompts Beginners Can Copy Directly

(See the prompts provided in Section 6 and 8 above for organizing resources, projects, topics, meetings, and weekly reviews.)

14. Consider Automation and Skills Later

Don't skip steps. Start with one-off tasks, then move to fixed prompts, then templates, and finally automation/Skills.

15. Troubleshooting

• Can't read base: Check if it's the same WeChat account.
• Poor quality: Be more specific in your prompts.
• 0 items shown: Check if files are in the base (not just dialog) and if parsing is complete.

16. Run One Minimal Loop First

Build the 3 bases, put in 20 items, run a review, and check the value. If it reduces your burden, expand.

Conclusion: A second brain isn't about having more data; it's about having a process to handle, filter, and judge that data.

About the Author

@AdrianPunk115