11. Text Generation - Write content
22. Chat Interface - Conversations
33. Summarization - Condense text
44. Analysis - Extract insights
55. Translation - Multi-language
66. Q&A - Answer questions

1UI Elements:
2- Input: Topic/keywords
3- Dropdown: Tone (professional, casual, fun)
4- Dropdown: Length (short, medium, long)
5- Button: Generate
6- Text: Result

1System:
2"You are a content writer. Generate content based on user input.
3Tone: {tone}
4Length: {length} (short=50 words, medium=150, long=300)"
5 
6User:
7"Write about: {topic}"

11. Button clicked
22. Validate inputs (not empty)
33. Call OpenAI API
4   - Build prompt with parameters
54. Display result in Text
65. Enable "Copy" button

1┌─────────────────────────┐
2│  Chat with AI           │
3├─────────────────────────┤
4│                         │
5│  [Bot] Hello! How can   │
6│        I help?          │
7│                         │
8│       [User] What is    │
9│              AI?        │
10│                         │
11│  [Bot] AI stands for... │
12│                         │
13├─────────────────────────┤
14│ [Type message...] [→]   │
15└─────────────────────────┘

1Message (Data Type):
2- content (text)
3- role (text: user/assistant)  
4- conversation_id (text)
5- created_date (date)
6- user (User)

1Page State:
2- conversation_id (generate unique ID on load)
3 
4Repeating Group:
5- Type: Message
6- Filter: conversation_id = Page's conversation_id
7- Sort: created_date ascending
8 
9Send Workflow:
101. Create Message (role: user, content: input)
112. Build messages array for API
123. Call OpenAI with full conversation
134. Create Message (role: assistant, content: response)
145. Clear input

1{
2  "model": "gpt-4o-mini",
3  "messages": [
4    {"role": "system", "content": "You are helpful assistant"},
5    {"role": "user", "content": "What is AI?"},
6    {"role": "assistant", "content": "AI is..."},
7    {"role": "user", "content": "Give me examples"}
8  ]
9}

1- Multiline Input (paste long text)
2- Dropdown: Summary type
3  - Brief (1-2 sentences)
4  - Key points (bullet)
5  - Detailed (paragraph)
6- Button: Summarize
7- Text: Summary result

1Brief:
2"Summarize this in 1-2 sentences: {text}"
3 
4Key Points:
5"Extract 5 key points as bullet points: {text}"
6 
7Detailed:
8"Write a comprehensive summary (150 words): {text}"

1Prompt:
2"Analyze the sentiment of this text.
3Return JSON format:
4{
5  'sentiment': 'positive/negative/neutral',
6  'score': 0.0-1.0,
7  'keywords': ['word1', 'word2'],
8  'summary': 'brief explanation'
9}
10 
11Text: {user_input}"

1In Bubble:
21. API response is text
32. Use "JSON Safe" plugin
43. Extract: response:formatted as JSON:sentiment

1Visual representation:
2- Positive: Green badge
3- Negative: Red badge
4- Neutral: Gray badge
5 
6Score: Progress bar (0-100%)
7Keywords: Tag chips

1As user types:
21. Detect pause (debounce)
32. Call AI with partial input
43. Suggest completion
54. User accepts or continues

1"Help me write" button:
2 
3Prompt:
4"User is filling a form field: {field_name}
5Current input: {partial_text}
6Suggest a complete, professional response."

1Before submit:
21. AI reviews form content
32. Suggests improvements
43. Checks for errors
54. User confirms or edits

1UI:
2- Input text
3- Source language (auto-detect)
4- Target language dropdown
5- Translate button
6- Result text

1"Translate the following text to {target_language}.
2Keep the tone and style.
3If there are cultural references, adapt them appropriately.
4 
5Text: {input_text}"

1Don't call AI for same input:
2 
31. Hash the prompt
42. Check database for existing response
53. If exists, return cached
64. If not, call AI and save

1For long responses:
2- Stream tokens as they generate
3- Better UX for chat
4- Requires server-side setup

1Good UX:
2- Disable button while loading
3- Show spinner/animation
4- "AI is thinking..."
5- Progress indicator