Skip to main content

​
Golden Rules for Effective Use

​
1. Clearly Describe Your Goals, Not the Process

❌ Bad Prompt: 
First use pandas to read the CSV,
then use matplotlib to draw the graph
✅ Good Prompt: 
Analyze this sales data, generate a
trend chart, and tell me the key findings
Happycapy automatically selects the best tools and methods. You just need to state the goal.

​
2. Provide Sufficient Context

❌ Bad Prompt: 
Help me fix this bug
✅ Good Prompt: 
There is no response after the user
clicks the login button.

Environment: React 18 + TypeScript
File: src/components/LoginForm.tsx
Browser console shows:
  'Cannot read property of undefined'

Expected behavior: After clicking,
an API request should be sent and
jump to the homepage

Please help me locate the problem
and fix it.

​
3. Work Iteratively, Don’t Ask for One-Time Perfection

❌ Bad Approach: 
Help me build a complete e-commerce
website, including user authentication,
product list, shopping cart, payment
integration, order management, review
system, recommendation algorithm...
✅ Good Approach: 
Help me build the basic framework first:
- Home page displays product list
- Product details page
- Simple shopping cart (no backend required)

After completion, we will gradually
add functions.
Principle: Break down big tasks → implement step by step → verify each step

​
4. Use Multi-Modal Capabilities

You can provide input in multiple formats:
  • Screenshots: Paste directly (Cmd/Ctrl + V)
  • Files: Upload PDF, Excel, Word, images
  • Web links: Share URLs for analysis
  • Audio/Video: Provide media for transcription and analysis

​
5. Save Successful Workflows

When a prompt works well:
  1. Record it (copy to notes)
  2. Mark usage scenarios
  3. Reuse or fine-tune directly next time
You can customize successful workflows as Skills, and all users can do this! We encourage you to create your own skill library to perform repetitive tasks with one click. Pro/Max users can also set up scheduled automation.

​
6. Regular Checkpoints to Facilitate Rollback

For long tasks or experimental work: 
Before implementing this function, let me confirm that
the current code has been saved.
Save strategies:
  • If using Git: commit regularly
  • If in Happycapy workspace: Download backups regularly
Think of it like saving in a game - easily roll back when something goes wrong.

​
7. Identify Your Skill Level

At the beginning of the conversation, state your experience level: 
I'm a designer and don't know much about programming.
Please explain in simple language.
Or: 
I am a senior developer, familiar with React and TypeScript.
You can give me code directly.
Happycapy will adjust communication style and level of detail accordingly.

​
8. Don’t be Afraid to “Start Over”

If the conversation goes astray:
Don’t try to correct the AI’s mistakes in a long thread. Directly open a new conversation and re-describe the task. It’s usually faster than correction.
Think of it like using a “slot machine”:
  1. Save state
  2. Try it for 30 minutes
  3. If satisfied, accept it; if not, try again

​
9. Use Happycapy’s Skill System

Happycapy supports Skills from the open source community, covering:

Image & Video

Generation and editing

Documents

PDF, Word, Excel, PPT

Web Development

Frontend and backend

Data Analysis

Visualization and ML

Research

Academic papers

Automation

Email, workflows
How to use Skills: 
Use ai-image-generation skill to help me generate a tech poster

Use comprehensive-researcher skill to research
'Latest Progress in Large Language Models'
View all available Skills by entering /help in the Happycapy dialog.

​
10. Max Plan Users: Make the Most of Your Extra Capacity

If you use Max Plan:
  • Take on ambitious, multi-step projects
  • Be bold and try complex tasks
  • Let Happycapy work autonomously for longer
  • Iterate multiple times until satisfied
That’s the value of Max Plan: significantly more room than Pro to get things done.

​
Common Misconceptions

Reality: AI is a collaborative partner, not a magic wand.Correct mindset:
  • Work iteratively
  • Provide feedback
  • Incremental improvement
Reality: Natural language is the best prompt.Correct method:
  • Describe your needs like chatting with a friend
  • No need to learn “Prompt Engineering”
  • Happycapy will understand your intentions
Reality: Happycapy is designed for non-technical people.Evidence:
  • Finance teams with zero programming experience execute complex data workflows
  • Legal teams with no development background build accessible applications in one hour
  • Product designers directly implement frontend functions

​
Quick Start by Role

​
Non-Technical People (3 Steps)

Try simple tasks: “Make me a poster”, “Analyze this Excel”
Save effective prompts and build your own workflow library
Set up automation (Pro/Max) to handle repetitive tasks

​
Developers (3 Steps)

Use Happycapy to understand unfamiliar code bases and generate tests
Let Happycapy implement small functions independently, you focus on reviewing
Build a self-validation loop to let Happycapy work autonomously for long periods

​
Data Analysts (3 Steps)

Upload data and let Happycapy do exploratory analysis
Let Happycapy build an interactive dashboard
Save common analysis workflows as reusable tools

​
Marketers (3 Steps)

Batch generate copy variations
Automated advertising creative production
Connect to MCP servers (Meta Ads, Google Ads) and build marketing automation

​
Product Designers (3 Steps)

Paste mockup and generate frontend code prototype
Implement simple style adjustments directly
Build a library of design system components

​
Productivity Comparison

Task TypeTraditional TimeHappycapy TimeTime Saved
Debugging complex infrastructure issues10-15 minutes~5 minutes50%
Generate advertising copy in batches2 hours15 minutes87%
Understand unfamiliar code basesHoursFew minutes80-90%
ML concept research and learning1 hour10-20 minutes80%
Create a data visualization dashboard1-2 daysHours75%
Global copywriting changes1 week1 hour95%
New employee onboarding code baseWeeksSeveral days70%
Design to frontend implementationBack and forth for several daysHours66%
Overall productivity improvement: 2-4x

​
Final Advice

​
Remember the Core Philosophy

  • AI should work for everyone - not just developers
  • Focus on “what you want” - not “how”
  • Max Plan — significantly more tokens than Pro
  • Chill like a capybara - relaxed, friendly, getting along with everyone

​
You Don’t Need To:

  • Learn the command line
  • Understand programming
  • Configure complex environments
  • Memorize prompt engineering tips
  • Worry about usage limits (Max Plan)

​
All You Need Is:

  • Say what you want to do
  • Trust Happycapy to choose the right tool
  • Work iteratively and gradually improve
  • Use AI to get more done

​
Need Help?

Join Discord

Connect with the community

Email Support

Get help from our team

Official Website

Learn more about Happycapy

Book a Walkthrough

Get a personalized demo
Happycapy, HappyYou Let AI become your most powerful assistant and return to the joy of work.