Mining demand from comments: 10 hits to 3 shootable topics

Pipeline

1. Search a batch of candidate posts
   Search 10 posts for a keyword and don't judge from the first one — the hottest is often a joke with no purchase intent.
2. Sort by interaction and take the top 3
   Add likes + comments + shares + views and sort on that. High interaction with few comments means it looks good but gives people nothing to talk about — skip it.
3. Pull comment insights post by post
   Three readings matter: question share (purchase or usage intent), top terms (the words people use for the pain), and top-liked comments (real tone you can lift straight into copy).
4. Roll it all up into one topic list
   Intersect the top terms that appear in all three — the intersection is the hardest pain under that term; then write headlines in the exact words of top-liked comments.

Real batch result

Endpoints you need

What this article uses get_comment_insights, with the parameters below; use REST when writing code and MCP when wiring a client — both return the same data.

Copy and run

curl -s https://lucydomi.com/mcp/v1/get_comment_insights \
  -H "Authorization: Bearer $BUSYBEE_API_KEY" \
  -H "Content-Type: application/json" \
  -X POST \
  -d '{"platform": "tiktok", "post_id": "7659488435199053088", "sample": 30}'

import os, json, urllib.request

KEY = os.environ["BUSYBEE_API_KEY"]
req = urllib.request.Request(
    "https://lucydomi.com/mcp/v1/get_comment_insights",
    data=json.dumps({"platform": "tiktok", "post_id": "7659488435199053088", "sample": 30}).encode(),
    headers={"Authorization": f"Bearer {KEY}", "Content-Type": "application/json"},
    method="POST",
)
data = json.load(urllib.request.urlopen(req, timeout=60))
for it in data["items"]:
    print(it.get("url"), it.get("text", "")[:60])

FAQ

Related reading