Imported from apidojo-io/apidojo-skills (skills/intent/extracting-tiktok-comments-for-research/SKILL.md). Install upstream with npx skills add apidojo-io/apidojo-skills --skill extracting-tiktok-comments-for-research. Copyright stays with the author (Apache-2.0).

Extracting TikTok Comments for Research

Pulls all public comments from TikTok videos for audience sentiment analysis, product research, or competitive intelligence. Comments are the rawest form of consumer voice — unfiltered reactions at scale.

Prerequisites

• APIFY_TOKEN environment variable set
• Optional: Apify MCP server installed

Inputs

Workflow

Progress:
- [ ] Step 1: Identify target video(s) and research goal
- [ ] Step 2: Run tiktok-comments-scraper
- [ ] Step 3: Fetch and clean comment dataset
- [ ] Step 4: Analyze themes, sentiment, and top comments
- [ ] Step 5: Deliver research output

Step 1: Clarify Parameters

Ask the user for:

• TikTok video URL(s) — direct links to specific videos (e.g., https://www.tiktok.com/@creator/video/[ID]) OR
• Creator handle — pull comments from their most recent/viral videos
• Max comments per video (default: 500; max: ~3,000)
• Research goal — sentiment analysis, product feedback, audience profiling, or competitive intel
• Date filter (optional — focus on recent comments only)

Tip for best research: Use 3-5 videos from the same creator or about the same topic for a reliable dataset.

Step 2: Run the Actor

Recommended — run_actor.js (handles waiting, output, and file saving automatically):

# Quick answer (prints table to chat)
node scripts/run_actor.js \
  --actor "apidojo~tiktok-comments-scraper" \
  --input '{"param": "value"}'

# Save as CSV
node scripts/run_actor.js \
  --actor "apidojo~tiktok-comments-scraper" \
  --input '{"param": "value"}' \
  --output YYYY-MM-DD_results.csv --format csv

# Save as JSON
node scripts/run_actor.js \
  --actor "apidojo~tiktok-comments-scraper" \
  --input '{"param": "value"}' \
  --output YYYY-MM-DD_results.json --format json

APIFY_TOKEN must be set in environment or .env file.

If Apify MCP is available:

Tool: apify:run-actor
Actor: "apidojo~tiktok-comments-scraper"
Input:
{
  "postURLs": [
    "https://www.tiktok.com/@[handle]/video/[VIDEO_ID_1]",
    "https://www.tiktok.com/@[handle]/video/[VIDEO_ID_2]"
  ],
  "maxCommentsPerPost": 500,
  "includeReplies": false
}

REST API fallback:

curl -X POST \
  "https://api.apify.com/v2/acts/apidojo~tiktok-comments-scraper/runs?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "postURLs": [
      "https://www.tiktok.com/@[handle]/video/[VIDEO_ID]"
    ],
    "maxCommentsPerPost": 500,
    "includeReplies": false
  }'

Wait for SUCCEEDED. Fetch dataset.

Step 3: Clean Comment Dataset

From raw dataset, extract per comment:

• text — the comment text
• author.uniqueId — commenter username
• diggCount — likes on the comment
• replyCommentTotal — how many replies this comment received
• createTime — timestamp

Clean:

• Remove empty or emoji-only comments (if doing text analysis)
• Remove spam patterns (repeated text, links, self-promotions)
• Remove the creator's own replies (identified by matching author.uniqueId)

Step 4: Analyze by Goal

Goal: Sentiment analysis Classify each comment as Positive / Negative / Neutral (use the same lexical method as the Twitter sentiment skill). Weight by diggCount — a liked comment reflects community agreement.

Goal: Product feedback Look for:

• Feature requests: "I wish", "you should", "would be better if", "needs"
• Pain points: "why doesn't it", "can't believe", "problem with", "doesn't work"
• Specific product mentions: nouns that repeat across multiple comments

Goal: Audience profiling From commenter bios (if available) and comment language:

• Identify audience demographics signals (age signals, geographic signals, interest signals)
• Find what questions the audience asks most

Goal: Top comments Simply sort by diggCount descending. Top-liked comments represent the community's most agreed-upon reactions.

Step 5: Format Output

Output Format

# TikTok Comment Analysis
Video(s): [N] | Total comments analyzed: [N] | Date: [DATE]

## Source Videos
| Video | Creator | Views | Comments Extracted |
|-------|---------|-------|-------------------|
| [url] | @[handle] | [N] | [N] |

## Sentiment Distribution (if goal = sentiment)
Positive: [X%] ([N] comments) | Negative: [X%] | Neutral: [X%]
Weighted by likes — Positive: [X%] | Negative: [X%]

## Top 10 Most-Liked Comments
| # | Comment | Likes | Replies |
|---|---------|-------|---------|
| 1 | "[comment text]" | [N] | [N] |

## Key Themes in Comments
| Theme | Frequency | Avg Likes per Comment |
|-------|-----------|----------------------|
| [Theme 1] | [N] | [N] |
| [Theme 2] | [N] | [N] |

## Most Asked Questions
1. "[question text]" — asked by [N] commenters
2. "[question text]" — [N] commenters

## Common Complaints / Pain Points
1. "[pain point]" — [N] comments, [N] total likes

## Audience Signals
- Age/demographic indicators: [summary]
- Geographic signals: [summary]
- Interest signals: [summary]

Troubleshooting

Few comments returned: Video may have comments disabled or be relatively new. Try a different video. All comments in non-English: Add a language filter post-processing, or adjust the search to English-language TikTok creators. Spam dominates results: Apply a filter: remove comments shorter than 5 words AND with 0 likes, which tend to be bots.