How to Get YouTube Transcripts in Make.com

Get a free API key

Create a key at transcript-yt.com. Every account gets 100 free credits, and one transcript is one credit. Keep the key out of individual modules where you can: store it once in a Make data store or a scenario variable and map it into the header.

Get one transcript with the HTTP module

Add the HTTP app's Make a request module. Set the URL to https://transcript-yt.com/v1/transcript and the method to GET, add an Authorization header with the value Bearer ts_live_..., and add a query string item url mapped to the video link from an earlier module, such as a Google Sheets row or an RSS item.

Turn on Parse response so the JSON comes back as fields you can map: title, author, language, segments, and text. If the next module only needs words, add format=text to the query string and map the raw data instead.

• URL: https://transcript-yt.com/v1/transcript
• Method: GET
• Headers: Authorization = Bearer ts_live_...
• Query string: url = the video link; optional language, translate_to, or format
• Parse response: Yes
• Timeout: raise it to 120 seconds or more, since videos without captions are transcribed from audio

Process a whole playlist or channel

Add a first HTTP module on GET https://transcript-yt.com/v1/playlist with url set to a playlist link or channel @handle. Listing is free and returns a videos array. Add an Iterator on videos, then the transcript module with url mapped to the iterator's videoId, and Make runs it once per video.

To use fewer operations, send up to 50 video IDs in one POST to /v1/transcripts with a JSON body of {"urls": [...]}, then iterate over the results array.

Send the transcript somewhere useful

Common next steps: add a row to Google Sheets with the title and text, create a Notion page per video, or pass the text to an OpenAI or Anthropic module with a prompt like "Summarize this transcript in 5 bullet points". The API can also return a summary itself: GET /v1/summary takes the same url parameter.

Make's free plan runs a scenario at most every 15 minutes, which is plenty for watching a channel's RSS feed and transcribing each new upload.

Handle errors

Add an error handler route to the transcript module so one bad video doesn't stop the scenario. Errors come back as {"error": {"code", "message"}}. Use a Resume or Break directive with retries for BLOCKED, RATE_LIMITED, and UPSTREAM_ERROR, and Ignore for VIDEO_UNAVAILABLE and NO_CAPTIONS. Failed requests are never billed.