All notable changes to this Actor are documented here. Dates are UTC.
Initial public release.
- Bulk TikTok video transcript extraction to WEBVTT.
- Single
urls input (array of one or more TikTok video URLs).
language selector across 8 languages (en, es, fr, de, it, ja, ko, zh).
use_ai_as_fallback boolean for videos without native captions (under ~2 minutes).
max_concurrency (1-5) for tuning parallelism.
- Per-row
_metadata (page, fetched_at, input_identifier).
- Per-row error isolation: failed items emit a row with
error set instead of crashing the run.
- Free plan cap: 50 results per run for free Apify users, uncapped for paying users.
- Retries with exponential backoff (1s, 2s, 4s) on transient errors (429, 502, 503, 504).
- Output schema, key-value store schema, and OpenAPI web server schema declared under
.actor/.
- Three dataset views: Overview, Transcripts only, Errors only.
- Live-view HTTP endpoints:
/, /status, /progress, /latest, /healthz.
- Acceptance-by-Use notice logged at run start.
- The Actor never finishes with status
FAILED. Configuration errors, missing input, fully invalid input, and unexpected runtime exceptions are all emitted as dataset rows and the run completes SUCCEEDED with a clear status message.