YouTube Transcript — FlashData

Read an existing transcript for timestamped notes. Defaults to English automatic captions; no ASR, translation or fallback to another caption track.

Install and connect

This is an independently packaged community Piece, not a confirmed official catalogue listing. It requires Activepieces 0.82.0+ and a Node.js 22.18+ worker. Your platform administrator must have custom Piece installation enabled; plan requirements vary.

Download the versioned npm package without running a data action:

npm pack '@flashdata-dev/[email protected]' --ignore-scripts

Upload the downloaded .tgz via Platform Admin → Catalogue → Pieces → Install Piece → Upload File. If your platform offers installation by package name, use @flashdata-dev/piece-youtube-transcript and version 0.1.0. Older platform versions may label the page Settings → My Pieces. Source build instructions are in the workspace guide.

1. Sign in or create a FlashData account.
2. Create an API Key with youtube_transcript permission. The secret is shown once; save it in an Activepieces connection.
3. Add YouTube Transcript — FlashData to a flow, choose YouTube Transcript, connect the key, and enter the source-specific fields. The source is fixed; this Piece cannot select another API.

Connection validation calls GET https://api.flashdata.dev/v2/api-access and checks this source permission. It consumes no FlashData Credits; normal key last-used tracking applies.

Example task

Start with a manual trigger and enter these example action inputs. Check the source-specific limits below before running the step.

{
  "video_id": "jNQXAC9IVRw",
  "language": "en",
  "auto": true
}

results[0].segments contains start, end and text; results[0].fullText contains joined text. The selected language and caption type must exist.

Map the returned data into your own table or research step. Search result content is data; review source links before relying on it.

Inputs

| Field | Type | Requirement | Constraint | | --- | --- | --- | --- | | video_id | string | Required | ^[A-Za-z0-9_-]{11}$ | | language | string | Optional | ^[A-Za-z0-9-]+$ | | auto | boolean | Optional | true / false or documented object |

Defaults and supported nested filter fields are documented in the source guide below. Explicit false values are preserved.

Cost, failures and limits

The connector code is free to install. FlashData data API calls are paid, using Credits; FlashData pricing applies. Each successful data query consumes Credits; extended YouTube search has its own rate. Activepieces execution and plan costs are separate. Clicking Test step executes the same paid operation as running it.

The connector does not automatically retry paid submissions. After a timeout, oversized response or uncertain server failure, check FlashData Jobs before resubmitting. Re-running a flow can still execute another paid query. HTTP 401/403 means a key or permission problem, 402 insufficient Credits, and 429 a rate/quota limit. Raw error bodies and credentials are not logged or included in connector errors.

Each HTTP read is limited to 65 seconds and 16 MiB. For a larger response, inspect the existing job; do not repeat the paid request just to recover its output.

YouTube Transcript inputs, outputs and limits · API reference · Privacy

Maintained by FlashData. This is not a Google- or YouTube-published integration.