Find TikTok Video Questions from Keywords Using Dumpling AI and GPT-4
Overview
This workflow automatically finds frequently asked questions from TikTok video comments based on keywords you provide. It uses Dumpling AI to search TikTok users and their videos, then applies GPT-4 to analyze the comments and extract top recurring viewer questions.
Key Benefits
- Instant content ideas: Discover what real viewers are asking about specific topics or niches.
- Data-driven strategy: Base your content calendar on actual audience questions, not guesses.
- Perfect for creators and marketers: Identify FAQs, engagement hooks, and viral opportunities fast.
- Fully automated: From keyword input to storing question results, the entire process runs hands-free.
Step 1: Receive Keyword Input
The workflow starts with a Form Trigger. When someone enters a keyword, this triggers the entire automation. For example, if you type “skincare,” it will search TikTok for relevant profiles and videos.
Why this matters: Keywords define your research scope. This makes the workflow flexible for any niche or topic.
Step 2: Search TikTok Users with Dumpling AI
The HTTP Request Node calls Dumpling AI’s TikTok user search API using the keyword.
It returns a list of TikTok accounts relevant to the keyword you entered.
Configuration details:
- Method: POST
- URL: https://app.dumplingai.com/api/v1/search-tiktok-users
- Body: query = {{ $json.Keywords }}
Why this matters: This step quickly finds active creators and influencers in your target niche.
Step 3: Split and Limit User Results
The Split Out Node breaks the list of returned users into individual items. Then, a Limit Node restricts the number of users processed (for example, to the first 3) to control API usage.
Why this matters: Limiting users helps you manage rate limits and keep the workflow efficient.
Step 4: Loop Through Each TikTok User
The Split In Batches Node processes each TikTok user one at a time. This looping structure is key for handling multiple users without overloading the APIs.
Why this matters: It ensures that the workflow remains stable when processing many accounts.
Step 5: Fetch TikTok Profile Videos
The HTTP Request Node fetches the latest videos for each TikTok user using Dumpling AI’s video scraping endpoint.
Why this matters: This gathers the real content you will later analyze for viewer questions.
Step 6: Split Videos for Processing
The Split Out Node takes the list of videos from each profile and splits them into individual video records, making it easier to process each video separately.
Why this matters: Analyzing videos individually ensures you capture all comment data accurately.
Step 7: Get Video Comments
For each video, the HTTP Request Node retrieves all comments using Dumpling AI’s TikTok comment API. It uses the video’s share URL as input.
Why this matters: Comments are where viewers express their questions and reactions, making this the core data for GPT-4 analysis.
Step 8: Extract and Clean Comments
The Python Code Node runs a small script to clean up the raw comment data. It extracts all comment text into a structured format, removing unnecessary metadata and nested replies.
Why this matters: Clean, structured comments help GPT-4 focus only on the relevant text, improving accuracy.
Step 9: Use GPT-4 to Find Top Viewer Questions
This is the intelligence step. The OpenAI Node with GPT-4 analyzes the cleaned comments to:
- Group similar questions.
- Identify the most frequently asked ones.
- Rank them clearly by frequency.
- Exclude spam and unrelated chatter.
Why this matters: GPT-4 transforms messy comment threads into actionable question lists that reflect real audience curiosity.
Step 10: Save Results to DataTable
Finally, the DataTable Node inserts the results into your TikTok Keywords table.
It stores:
- The TikTok video URL
- The extracted viewer questions
Why this matters: This gives you a searchable database of questions tied to videos, perfect for research, content planning, or reporting.
Tools and Accounts Needed
- Dumpling AI API key for TikTok search and comment extraction
- OpenAI GPT-4 API key for language processing
- n8n DataTable for storing extracted questions
Conclusion
This workflow turns keywords into curated lists of real TikTok viewer questions automatically.
It is perfect for:
- Content creators who want to make videos that answer real audience questions.
- Marketing teams researching trends and pain points.
- Agencies building FAQ content calendars for multiple clients.
- Strategists looking for viral engagement angles.
Once set up, you can simply submit keywords and let the workflow continuously find the most relevant questions for you.
Download the blueprint used in this blog post
Click here to access the blueprint.
