Pull TikTok post metadata, engagement counts, captions, and author details as structured JSON, built for analytics and archiving at scale.
If you are searching for a TikTok video API, you almost certainly need three things working together: a clean way to pull a creator's full video feed, structured data for any single video, and a keyword search that returns ranked videos. TikLiveAPI delivers all three from one host, https://api.tikliveapi.com, with a single header for authentication and JSON responses you can read without reverse-engineering anything. Every request is a plain GET, so you can prototype in a browser tab, wire it into a cron job, or call it from a serverless function with no SDK to install.
This page focuses on the video-centric endpoints in the posts category. They are the building blocks behind competitor monitoring dashboards, content archivers, trend-research tools, and no-watermark download flows. Credits are deducted per call by the API itself, they never expire, and new accounts start with 100 free credits so you can validate the responses against your own use case before spending a cent.
The /user-posts/ endpoint takes a numeric userid and returns a videos array. Each item carries the video_id, caption title, cover image, duration, the play URL, and the full stats block: play_count, digg_count, comment_count, share_count and collect_count. Page through an entire back catalog with the cursor field.
Pass a video URL to /post-detail/ and receive a single object with the same rich stats plus three media URLs: play and wmplay for the standard stream and watermarked copy, and hdplay for the high-definition version. The nested music_info object identifies the original sound behind the clip.
The /post-comments/ endpoint returns a comments array for any video URL. Each comment includes the text, its own digg_count, a reply_total, the create_time, and a full user object. Fetch up to 50 per call and walk the cursor to capture the whole thread for sentiment analysis.
The /search-video/ endpoint takes a keyword and returns ranked videos in the same shape as the user feed. Narrow results with publish_time (past 24 hours through last 6 months), reorder with sort_by (relevance, like count, or date posted), and target a market with the optional region code.
Both /post-detail/ and the dedicated /download-video/ endpoint expose direct play URLs that point to clean, watermark-free MP4 streams. These are the same links you can hand to a downloader or pipe into an archive job, with the watermarked variant available in wmplay when you need it.
Beyond single videos, the posts category fans videos out by source: /music-posts/ lists clips using a given sound, /challenge-posts/ streams videos behind a hashtag, and /playlist-posts/ returns an ordered series. All share the familiar videos shape with count and cursor pagination.
Authentication is one header, X-Api-Key. Here is a real, runnable request that fetches the most recent videos from the official TikTok account (user ID 107955):
curl "https://api.tikliveapi.com/user-posts/?userid=107955" \
-H "X-Api-Key: YOUR_API_KEY"
The response is a videos array. Trimmed to the fields that matter for analytics, a single item looks like this:
{
"videos": [
{
"video_id": "7571916663675833631",
"title": "POV: you foster one dog and accidentally start a nationwide movement",
"duration": 76,
"play_count": 221596,
"digg_count": 8166,
"comment_count": 507,
"share_count": 278,
"collect_count": 205,
"create_time": 1762974261,
"author": { "id": "107955", "unique_id": "tiktok", "nickname": "TikTok" }
}
]
}
Swap user-posts for search-video?keyword=cat and you get the same video shape for any keyword. Every field above is returned verbatim by the API, so your parsing code is identical across the user feed, search, music, hashtag, and playlist endpoints.
Poll /user-posts/ on a schedule for the accounts you track, store each video_id with its stats, and watch how play_count and digg_count move over time. This is the data spine behind any dashboard that monitors competitor TikTok accounts at scale.
Run /search-video/ with publish_time=1 and sort_by=1 to surface videos from the last 24 hours ranked by likes. It is a fast read on what is gaining traction for a keyword, and pairs naturally with hashtag tracking from /challenge-posts/.
Resolve any video URL through /post-detail/, grab the clean play or hdplay link, and download it for a content archive or compliance record. Bulk archive pipelines combine this with playlist and user-feed enumeration to capture entire catalogs.
Stream a video's full discussion with /post-comments/, page through the cursor, and feed the text values into a sentiment or summarization pipeline. The per-comment digg_count and reply_total let you weight the loudest voices.
Create an account and you get 100 credits to spend across any of these video endpoints, no card required. Test live in the playground, copy the generated code snippet, and ship. Credits are pay-as-you-go and never expire, so you only pay for the calls you actually make.