YouTube transcript in Zapier

Option 1: Webhooks by Zapier (GET)

Add an action step, choose Webhooks by Zapier, event GET. Webhooks is a premium app, so it needs a paid Zapier plan (as of September 2026).

Test the step. The API returns JSON with a transcript array whose items carry start and duration in seconds and the text:

{
  "video_id": "dQw4w9WgXcQ",
  "transcript": [
    { "start": 0.0,  "duration": 3.52, "text": "We're no strangers to love" },
    { "start": 3.52, "duration": 2.88, "text": "You know the rules and so do I" }
  ]
}

Zapier shows the array as line items:

video_id            dQw4w9WgXcQ
transcript          (line items)
transcript[]start   0, 3.52, 6.4, ...
transcript[]text    We're no strangers to love, You know the rules and so do I, ...

To turn that into one paragraph without code, add Formatter by Zapier → Utilities → Line-item to Text, input Transcript Text, separator a single space. Its output is the plain transcript for the next step. The full parameter list for the endpoint is in the API docs.

Option 2: Code by Zapier (JavaScript)

One step that does all three jobs — extract the id, fetch the transcript, join the segments — and works on every plan. Add Code by Zapier → Run JavaScript, map two Input Data fields, paste the script:

// Code by Zapier · JavaScript
// Input Data:  url  ←  1. Video URL        (from the trigger)
//              key  ←  ta_your_key         (typed in once, stored by Zapier)
const m = String(inputData.url).match(
  /(?:v=|youtu\.be\/|\/shorts\/|\/embed\/|\/live\/)([A-Za-z0-9_-]{11})/
);
if (!m) throw new Error('No YouTube video id in: ' + inputData.url);
const videoId = m[1];

const res = await fetch(
  'https://api.transcriptapi.io/transcript?video_id=' + videoId + '&language=en',
  { headers: { Authorization: 'Bearer ' + inputData.key } }
);
const body = await res.json();

if (!res.ok) {
  // 404: captions off / private / deleted / age-restricted. Refunded.
  // retryable:false means the video is the problem, so end the Zap quietly.
  if (body.retryable === false) return { videoId, text: '', skipped: body.detail };
  // 402 (no credits) or 502/503 (YouTube refused, refunded): let Zapier
  // mark the run as errored and replay it later.
  throw new Error(res.status + ': ' + body.detail);
}

const text = body.transcript
  .map((s) => s.text.trim())
  .filter(Boolean)
  .join(' ');
return { videoId, segments: body.transcript.length, text };

Output fields text, videoId and segments become mappable in later steps. A one-hour talk is around 9,000 words of text; the fetch takes a second or two on a cache miss and well under that on a hit.

Getting the 11-character id

The YouTube trigger gives you a Video Id field directly; map that and skip extraction. Anything else — a form, a Slack message, a Notion property, an RSS item — gives a URL, and video_id wants only the id. The regular expression in the Code step handles watch?v=, youtu.be/, Shorts, embed and live links. Without Code, use Formatter → Text → Extract Pattern with (?:v=|youtu\.be/|shorts/)([A-Za-z0-9_-]{11}).

Send a full URL, or a field that never resolved, and the API stops you before it spends anything:

HTTP 422
{
  "detail": "video_id still contains an unsubstituted template placeholder ('{{123456__id}}'). Your automation sent the variable rather than its value — check the expression that builds this request."
}

That second case happens when a Zap is tested before the trigger has a sample record: Zapier sends the raw field token. Pull in a real sample, re-test, and the token becomes an id. 422 is never charged.

The Zap: new channel video → summary in Slack or Notion

1. YouTube · New Video in Channel        (trigger, polls the channel)
2. Code by Zapier · Run JavaScript        (id extraction + transcript + join)
3. Filter by Zapier · only continue if    Text · (Text) Exists
4. ChatGPT · Conversation                 (or Anthropic · Send Message)
   "Summarise in five bullets with the timestamp of each point:"
   + Step 2 · Text
5. Slack · Send Channel Message           (or Notion · Create Page)

Step 3 matters: a video with captions off comes out of the Code step with an empty text, and the filter stops the Zap instead of sending an empty summary to the channel. Notion's page body is capped at 2,000 characters per block, which is fine for a summary and wrong for a whole transcript — post the summary and a watch link, not the text.

The YouTube trigger polls on Zapier's schedule for your plan. If you would rather be told, POST /channel/subscribe with a Webhooks by Zapier → Catch Hook URL delivers {channel_id, video_id} to the Zap within 15 minutes of an upload, free. Details and a FastAPI receiver are in the cookbook.

Failures, refunds and replays

Captions off, private, deleted or age-restricted videos have nothing to return. They come back as 404, the credit is refunded, and retryable is false:

HTTP 404
{
  "detail": "This video has captions turned off, so there is no transcript to fetch. Your credits have been refunded.",
  "error": "TranscriptsDisabled",
  "hint": "Retrying will not help — the uploader has disabled captions for this video.",
  "retryable": false,
  "credits_refunded": true
}

A Webhooks step treats any 4xx or 5xx as an error and halts the Zap, which is the right behaviour for 402 (balance empty) and 502/503 (YouTube refused, refunded, replay it from the Zap history in a few minutes). It is noisy for 404, which no replay will fix. The Code step above handles that: it returns an empty text on retryable: false and throws on everything else, so only failures worth replaying show up as errors.

Brand-new uploads are the special case: auto-generated captions can lag the video by minutes to hours. A Delay by Zapier step of 20 to 30 minutes after the trigger avoids most of those first-attempt NoTranscriptFound responses.

Cost per run

Zapier: one task per action step, so the outline above is four tasks per video (Filter counts, the trigger does not). TranscriptAPI: one credit per transcript. A cached repeat of the same video is free, a refunded failure is free, the channel webhook is free. Translation adds one credit per 40 segments, minimum two.

A channel that uploads daily uses about 30 credits a month. Accounts start with 20 free credits and no card; 2,500 credits are a $10 one-time top-up that never expires; the Production plan is $49 a month for 25,000. Full table on the pricing page.

What we do not do: download the video or audio, return comments or view counts, or manage a channel you own. If there is no caption track, there is no transcript, and you are not charged for finding that out.

Frequently asked questions

The same integration in n8n and Make.com; in code, see LangChain. If an AI assistant should read the video directly, the MCP server needs no Zap at all. Get a free key with 20 credits at /login.