Superbash
SearchUser Hub

Decodo Web Scraping · Tutorial 07

Research Reddit and TikTok without losing context

Collect public posts, community activity, and engagement fields while preserving URLs, timestamps, scope, and the difference between observation and fact.

Decodo dashboard showing Reddit Post and TikTok templates among specialized web scraping targets.
Reading time
10 min
Last updated
August 2026

0 of 11 complete

Complete & next →

Decodo’s social tools help collect public content and engagement fields. Your assistant still must distinguish what a person posted from what is independently true.

Current social tools used here

  • reddit_post — one Reddit post and its available discussion data.
  • reddit_subreddit — posts from a named subreddit.
  • reddit_user — a public user’s profile and available activity.
  • tiktok_post — one public TikTok post with caption, hashtags, and engagement fields.

TikTok Shop belongs to the ecommerce tutorial. YouTube gets its own tutorial next.

Decodo dashboard target catalog with Reddit Post, TikTok, search, ecommerce, and AI scraper categories.
Target-specific templates reduce parsing work, but the content still needs editorial judgment.

Workflow A: inspect one Reddit post

Use Decodo's reddit_post tool on [PUBLIC REDDIT POST URL].
Return: post title, subreddit, author name as displayed, posted time, score,
comment count, post body, outbound URL, post URL, and the five highest-ranked
available comments with comment URLs/IDs. Use one request.

Review the post date. A high score does not make a claim accurate.

Workflow B: scan a subreddit

Use reddit_subreddit for r/[NAME].
Collect the top 20 posts from the past week.
Return title, author, posted time, score, comment count, flair, and URL.
Then group recurring questions by topic. Keep the raw post list before synthesis.
Use one request.

State the sorting rule and time window. “Popular topics” means nothing without both.

Workflow C: inspect public user activity

Use reddit_user only for a legitimate task and minimize personal data. Do not assemble identity profiles, infer sensitive traits, or connect pseudonymous accounts to real people.

Use reddit_user on [PUBLIC PROFILE URL].
Return the 10 most recent public posts/comments with subreddit, timestamp,
type, short excerpt, and permalink. Do not infer identity or personal traits.
Use one request.

Workflow D: capture one TikTok post

Use tiktok_post on [PUBLIC TIKTOK POST URL].
Return creator handle, caption, hashtags, posted time, duration, view count,
like count, comment count, share count, music/audio fields if available,
canonical URL, and observed time. Set xhr true only if needed.
Use one request.

Engagement counts are a snapshot. Label them with observation time.

Turn posts into a beginner-friendly brief

Keep three layers separate:

  1. Raw records: exact URL, timestamp, author/handle, text, engagement.
  2. Observed patterns: repeated questions or themes in this collected set.
  3. Verified claims: statements checked against primary or authoritative sources.

Useful output:

Observation: 8 of 20 collected posts ask about battery life.
Community claim: several commenters say update 2.1 reduced runtime.
Verification status: not yet checked against release notes or controlled tests.

Stop conditions

Stop when:

  • the content is private, deleted, age-gated, or login-only;
  • the task asks for sensitive profiling or harassment;
  • geography or account state would misrepresent what was observed;
  • repeated retries exceed the stated budget;
  • a post URL no longer resolves to the expected content.

Official references