TikTok Scraper (Search, User Posts & User Stories)

Scrape TikTok videos by keyword/hashtag search, or pull every post/story from one or more usernames — no TikTok login required for search and public posts.

This Actor is an Apify wrapper around the tiktok-end-to-end-crawler engine: the same async TikTokAPI client (pagination, retries, rate-limiting) that already powers the original CLI tool / Kafka+Elasticsearch pipeline, just re-pointed at the Apify dataset and Apify Proxy instead of Kafka/ES/argparse.

Input

Field	Type	Notes
operation	search | userPosts | userStory	What to scrape
searchQuery	string	Required for search
usernames	array of strings	Required for userPosts / userStory. Multiple usernames run in one job.
resultsPerPage	int (1-50)	Items requested per page, default 12
maxPagesPerQuery	int	Pages to paginate per keyword/username, default 1
maxItems	int	Hard stop for the whole run, default 50 (0 = unlimited)
hd	bool	Request HD video links, default true
proxyConfiguration	object	Apify Proxy — recommended, the upstream API rate-limits by IP
cookies	string (secret)	Only needed if userPosts/userStory return HTTP 403
rateLimitRps	int	Requests/second cap for this run, default 5

Output (one dataset row per video)

Each row is the full TikTokPost model (id, title, description, author, music, video, stats, hashtags) plus a few flattened convenience fields: authorUsername, playUrlNoWatermark, playUrlWatermark, diggCount, playCount, commentCount, shareCount, operation, queryValue, scrapedAt.

What changed vs. the original CLI project

• Dropped: Kafka/Elasticsearch output drivers, the Controllers/Input/ Output abstraction layer, env/YAML settings loading. Apify's dataset + input schema already cover that role, so keeping them would just add unused dependencies (slower builds = more compute cost on every run).
• Kept as-is: TikTokAPI (tiktok_api.py) and the Pydantic schemas (schemas.py) — same pagination/retry/backoff logic, same field aliases, same model_dump(by_alias=True, exclude_none=True) shape. KafkaEvent was renamed ScrapeEvent since there's no Kafka here.
• Added: Apify Proxy wiring (proxy_url was already a config field in the original CrawlerSettings — it just wasn't connected to anything before), and PPE charging via Actor.push_data(item, charged_event_name=...).

Monetizing this Actor

1. Push this folder to a GitHub repo (or use apify push from the Apify CLI).
2. In Apify Console → Actor → Monetization, set up Pay per event (PPE) and define an event named post-scraped (matches CHARGED_EVENT_NAME in src/main.py) — e.g. price it somewhere in the $1–10 per 1,000 results range, in line with typical Store pricing.
3. You don't need to touch the code further — Actor.push_data(..., charged_event_name="post-scraped") only actually charges once that event exists in Console; until then it's a no-op.

Possible follow-ups (not built into v1, to keep the build light)

• Optional second output: also forward each item to your own Kafka topic or Elasticsearch index (you already have working drivers for both in the original project's helpers/output/driver/) — could be pitched as an "enterprise" / advanced-input toggle on top of the free dataset output.
• A 4th postDetail operation that re-uses fetch_post() to grab a single video by ID/URL.