Bilibili Transcript API
Turn any B站 video into text — handles long-form content, BV/av IDs, and b23.tv links.
Bilibili hosts long-form Chinese video — lectures, tech reviews, documentaries — where a transcript is worth the most. This API accepts BV IDs, legacy av IDs, full URLs, and b23.tv short links, and returns the complete spoken content as timestamped text.
Long videos are handled natively: a 30-minute lecture comes back as one clean transcript with sentence-level timestamps, ready for summarization, search indexing, or subtitle files.
Languages: Mandarin Chinese (best-in-class), English, and 6 more
Call it from code
The API runs on Apify's platform — one HTTP call starts a run and returns the dataset. Get a free API token by signing up (includes $5/month free usage).
# curl — synchronous run, returns transcript items directly
curl -X POST \
"https://api.apify.com/v2/acts/ethereal_wool~bilibili-transcript-scraper/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
-H "Content-Type: application/json" \
-d '{"videoUrls": ["https://www.bilibili.com/video/BV1GJ411x7h7"]}'# Python
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("ethereal_wool~bilibili-transcript-scraper").call(run_input={
"videoUrls": ["https://www.bilibili.com/video/BV1GJ411x7h7"],
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
print(item["transcript"])
for s in item["sentences"]:
print(s["beginTime"], s["text"])Pricing
No subscription, no minimum. Runs that return nothing are never billed.
Frequently asked questions
Does it work on long videos?
Yes — there is no hard duration cap. A 30-minute video simply bills 30 × $0.025 = $0.75. Billing is strictly per minute of audio transcribed.
Which URL formats are accepted?
BV IDs (BV1xx…), legacy av IDs, full bilibili.com/video URLs, and b23.tv share links.
Multi-part (分P) videos?
The first part is transcribed by default; pass the URL with ?p=N to transcribe a specific part.
Why not just use Bilibili's own subtitles?
Most Bilibili videos have no CC subtitles. When they exist they're often auto-generated with poor accuracy. This API transcribes the audio directly with a stronger Mandarin model.
What's in the response?
Full transcript text, timestamped sentences, plus video metadata: title, uploader, play/danmaku counts, duration.
Other platforms
TikTok
Turn any TikTok video URL into clean, timestamped text with real AI speech recognition.
Douyin (抖音)
Turn any Douyin (抖音) video into text — best-in-class Mandarin speech recognition.
Xiaohongshu (小红书 / RedNote)
Turn any 小红书 video note into text — Mandarin-optimized speech recognition.
Instagram Reels
Turn any Instagram Reel into text with real AI speech recognition.
YouTube
Fetch YouTube transcripts at scale — by URL, video ID, or search keyword.
Any URL
One endpoint for TikTok, Douyin, Xiaohongshu, Bilibili, and Instagram Reels — paste any link, get text.