Setting Up the YouTube API Request in Make.com

Use Make.com’s HTTP module to query the YouTube Data API (search.list) with the following parameters:

• q=[keyword] → Searches for videos containing a specific keyword

• type=video → Returns only videos (not playlists or channels)

• part=snippet → Fetches video metadata (title, description, channel, etc.)‍

• maxResults=50 → Retrieves the top 50 results

• resultsorder=viewCount → Sorts videos by most views

• publishedAfter=7 days ago → Fetches only the latest videos

Expanding the Search Query with Multiple Variations

Since each YouTube API request only returns 50 results, you can increased the dataset by running multiple keyword searches simultaneously. For example, when searching for cover songs, I used the following keywords:‍

• "cover"

• "cover song"

• "acoustic cover"

• "guitar cover"

• "live cover"

• "piano cover"

This method can be applied to any topic, replacing "cover" with keywords like "AI tools", "travel vlog", or "fitness workout".

‍How to Implemented Dynamic Queries in Make.com‍

To loop through multiple queries dynamically, place a Set Variable module at the beginning of the scenario in Make.com with this value:

‍{{add(emptyarray; "cover"; "cover song"; "acoustic cover"; "guitar cover"; "live cover"; "piano cover")}}

Then, add an Iterator to loop through each search term and send a separate API request, and connect it to the existing http module.

‍Result:
This retrieves a larger, more diverse set of videos uploaded in the past week.

Some of the fetched videos weren’t in English, so I used LangDetect API in Make.com to identify the video’s primary language:

Create an HTTP module in Make.com

This module sends a request to the LangDetect API with each video title as input.

Configure the HTTP Request

• Method: POST
• URL: https://api.langdetect.com/detect
• Headers: Include the necessary API key for authentication.
• Body: Send the video title as a JSON payload, for example:
{ "q": "Shape of You cover song live performance" }

Extract the detected language

The API response returns a language code (e.g., en for English, es for Spanish). Create a filter and extract the language field from the response.

Apply a confidence threshold

The API also provides a confidence score indicating the accuracy of the language detection. Set a threshold (e.g., 4) to filter out low-confidence detections.

Handling Edge Cases

Some video titles contain mixed languages, leading to false positives. To mitigate this, keep videos where English appears as a detected language but review those with lower confidence scores.

If a video is mistakenly flagged as non-English, it can still be manually reviewed in Airtable before final approval.

Result:
‍• If the detected language is en and the confidence score meets the threshold, keep it.
• If the detected language is not en, exclude it from the dataset.

This filtering process ensures that only high-confidence English-language videos remain, improving the relevance of the dataset.

With relevant videos identified, the next step is to extract additional details about each video and its channel using Make.com.

Fetching Video Details

To retrieve more information about each video, set up an HTTP module in Make.com and use the YouTube videos.list API:

• Method: GET
• URL: https://www.googleapis.com/youtube/v3/videos
• Parameters:
-- part=snippet,contentDetails,statistics
-- id=[video_id]
-- key=[YOUR_API_KEY]

This API call provides details such as video duration, view count, category, etc.

Fetching Channel Details

To retrieve more information about each channel, set up another HTTP module in Make.com and use the YouTube channels.list API:

• Method: GET
• URL: https://www.googleapis.com/youtube/v3/channels
• Parameters:
-- part=snippet, statistics
-- id=[channel_id]
-- key=[YOUR_API_KEY]

This API call provides details such as subscriber count, view count, and video count.

Now that all the relevant video and channel data has been gathered, the next step is to store this information in Airtable for easy access and organization.

Create a new Airtable base and define the following fields:

Use the Create Record module in Make.com to send each processed video’s data to Airtable. Map the API response fields to the corresponding Airtable fields.

Result:
Our Make.com scenario is now complete. All of our video data is now structured, searchable, and easy to manage, allowing for seamless trend tracking and filtering.

While all of our video data is now in Airtable, there are a couple steps that we can take to enhance the readability of our results.

Video Title Formatting Fix

In Airtable, video titles are sometimes imported like this:

'Your Love' By The Outfield | Kelly Clarkson Kellyoke Cover

To fix the formatting, add this formula to the primary Video Title formula field:

SUBSTITUTE(
  SUBSTITUTE(
    SUBSTITUTE(
      SUBSTITUTE(
        {Raw Video Title}, 
        "'", "'"
      ),
      """, '"'
    ),
    "&", "&"
  ),
  "&lt;", "<"
)

Result:
Now the primary field of the table will display clean easy to read video titles:

'Your Love' By The Outfield | Kelly Clarkson Kellyoke Cover

Displaying Video Thumbnails in Airtable

To enhance the visual presentation of the data, we can display video thumbnails directly in Airtable rather than just storing the URL.

Create an attachment field in Airtable named Thumbnail. Then in the make.com Create A Record module, map the File URL field of the attachment field to the thumbnail URL from the API call.

Initially, I filtered the cover song query using videoCategoryId=10 (Music category), which restricted results to only return videos uploaded under the Music category. However, I found that this was too restrictive, as some covers were uploaded into different categories (e.g., Entertainment or People & Blogs).

To capture a broader range of relevant videos, I removed the category restriction to include all categories. Then, I set up filters in Airtable to manage the results more effectively.

Create a formula field named Category, and use this formula to translate the Category IDs from the Video Category ID column (CategoryID) into their category names:

I also wanted to be able to filter out YouTube Shorts.

Unfortunately it’s not possible to filter shorts out of the initial query, so I used an Airtable formula to convert the duration from the Video Duration column (Duration) from PT2M50S format to seconds:

IF(
  FIND("M", {Duration}),
  VALUE(MID({Duration}, 3, FIND("M", {Duration}) - 3)) * 60,
  0
) +
IF(
  FIND("S", {Duration}),
  VALUE(MID({Duration}, FIND("M", {Duration}) + 1, FIND("S", {Duration}) - FIND("M", {Duration}) - 1)),
  0
)

Then, I added a Shorts column in Airtable with this formula:

IF({Duration (Seconds)} <= 60, "Yes", "No")

Result:
Now, we can easily filter out shorts or create a specific shorts Airtable view.

Since we are doing a new search each week, the ability to view each week's search results separately will be helpful when reviewing the results.

Adding a Created field in Airtable provides the date that each search result enters the database. With a formula, we can use it to create a field called Which Week:

Now that all of the data is neatly organized and filterable in Airtable, we can use filters to create a review process for each week's data search.

I chose to use a checkbox column called Approved to manually approve the videos in the list each week. I set up three Airtable views to aid the review of each week's data:

High Confidence – This week's videos that are most likely to be relevant.
Video category is Music, Videos are not Shorts, High Language Confidence.
‍
Low Confidence – This week's videos that could be relevant.
Other video categories, Videos are not Shorts, Medium Language Confidence.
‍
No Confidence – This week's videos that should be excluded.
Shorts, Videos with specific keywords (e.g., reaction, vlog), Low Language Confidence.

To view the final list, I created a new view called Approved. Creating a grouping using the Created field provides a clear view of all of the approved content week by week.

Lastly, we can use Airtable's Interface feature to take the list of approved content from each week and format it into a comprehensible, interactive, and shareable format.

Top Cover Song videos uploaded each week  

This dashboard collects the 50 most viewed videos uploaded in each week for the following keywords:
‍#cover, #cover song, #acoustic cover, #live cover, #guitar cover, #piano cover

YouTube Shorts, non-English, and irrelevant content are removed. Search refreshes every Sunday at 6am PST.

This Week

Last Week

Previous Weeks

What are s-views, s-likes, and s-subs?
These "S" metrics are obtained for each video at the time of each weekly search (Sundays, 6am PST). Because some videos are uploaded earlier in the week than others, not every video has had equal time to perform.

Don't see a specific video?
Videos where creators did not use any of the specified search keywords will not appear in this list

Results:
The end product is an automated, shareable, easy to read report on the top performing videos for any keyword each week on YouTube.

Now, every week:

The top trending videos for any keyword are fetched automatically
Non-English videos are filtered out
Video thumbnails are displayed as images in Airtable
Shorts are flagged & removed
Videos are filterable by category
Data is auto-organized by week