11. OpenAI account (platform.openai.com)
22. API Key (Settings → API Keys)
33. Credits/payment method
44. API endpoint knowledge

1URL: https://api.openai.com/v1/chat/completions
2 
3Method: POST
4 
5Headers:
6- Authorization: Bearer sk-your-api-key
7- Content-Type: application/json
8 
9Body:
10{
11  "model": "gpt-4o-mini",
12  "messages": [
13    {"role": "system", "content": "You are helpful"},
14    {"role": "user", "content": "Hello!"}
15  ],
16  "max_tokens": 500
17}

11. Plugins → API Connector
22. Add new API:
3   - Name: OpenAI
4   - Authentication: Private key in header
5   - Key name: Authorization
6   - Key value: Bearer sk-...
7 
83. Add API Call:
9   - Name: Chat Completion
10   - Method: POST
11   - URL: https://api.openai.com/v1/chat/completions

1{
2  "model": "gpt-4o-mini",
3  "messages": [
4    {
5      "role": "system", 
6      "content": "<system_prompt>"
7    },
8    {
9      "role": "user",
10      "content": "<user_message>"
11    }
12  ],
13  "max_tokens": <max_tokens>
14}

11. Click "Initialize call"
22. Enter test values
33. Bubble detects response structure
44. Save API

1Workflow:
2When: Button clicked
3Action: OpenAI - Chat Completion
4  - system_prompt: "You are a helpful assistant"
5  - user_message: Input's value
6  - max_tokens: 500
7 
8Then: Display result in Text element
9  - Text: Result of step 1's choices:first item's message content

1Pre-built plugins (easier):
2- "OpenAI GPT" plugin
3- "AI Text Generator"
4 
5Install → Configure API key → Use actions

1Glide cannot call APIs directly.
2Use Make (Integromat) as middleware:
3 
4Glide → Make Webhook → OpenAI → Return to Glide

11. Create Make scenario
22. Webhook trigger (Custom webhook)
33. Copy webhook URL
44. Add OpenAI module:
5   - Create a Completion
6   - Model: gpt-4o-mini
7   - Prompt: {{data from webhook}}
85. Webhook Response module
96. Return AI response

11. Button action → Call API
22. URL: Make webhook URL
33. Method: POST
44. Body: {"prompt": column value}
55. Save response to column

1Glide has built-in AI column:
21. Add column → AI
32. Configure:
4   - Input: Product Description
5   - Task: "Summarize this product in 1 sentence"
63. AI auto-generates for each row

11. API Calls tab
22. Add API Call:
3   - Name: OpenAI Chat
4   - Method: POST
5   - URL: https://api.openai.com/v1/chat/completions
6   
73. Headers:
8   - Authorization: Bearer [API_KEY]
9   - Content-Type: application/json
10   
114. Body (JSON):
12   {
13     "model": "gpt-4o-mini",
14     "messages": [
15       {"role": "user", "content": "[user_input]"}
16     ]
17   }
18   
195. Variables:
20   - user_input (String)

11. Test API call
22. Define response schema:
3   - choices[0].message.content
4 
53. Use in widget:
6   - Text widget
7   - Value: API response → content

1Button pressed:
21. Set loading state = true
32. Call OpenAI API
4   - user_input: TextField value
53. Store response in Page State
64. Set loading state = false
75. Display response

11. Set max_tokens limit
22. Use gpt-4o-mini (cheaper)
33. Cache responses when possible
44. Rate limit user requests
55. Monitor usage dashboard

1Handle cases:
2- API timeout
3- Invalid response
4- Rate limit exceeded
5- Insufficient credits
6 
7UI feedback:
8- Loading states
9- Error messages
10- Retry options

1Page elements:
2- Input (topic)
3- Dropdown (tone: formal/casual)
4- Button (Generate)
5- Text (result)
6- Loading indicator
7 
8Workflow:
91. Button clicked
102. Show loading
113. Call OpenAI API:
12   - system: "Write in {tone} tone"
13   - user: "Write about {topic}"
144. Hide loading
155. Display result

1System: "You are a professional content writer. 
2Write in {tone} tone. Be concise and engaging."
3 
4User: "Write a short paragraph about: {topic}"