Review mining with AI: what buyers praise and complain about
Your AI reads review videos' comments and shop reviews of a product, and tells you what buyers praise, what they complain about and how often, with quotes. It uses the Comment analysis tool and 6 interfaces. A real run recorded 2026-10-18 answered “What do viewers complain about in the comments of new Kindle reviews on YouTube?” with 1 request for 50 credits (≈ $0.10–$0.17). Failed calls are free.
What you get
Your AI reads review videos' comments and shop reviews of a product, and tells you what buyers praise, what they complain about and how often, with quotes.
Reviews are spread over shops and review videos, and the comments under a review often say more than the review. Reading enough of both to count what comes up is slow.
What it uses
The ready-made tool does the whole job in one request: SocialToAI runs the searches, reads the posts and comments, and returns one result with its sources. Each is priced before it runs.
Your AI can also call the interfaces one at a time. These are the ones this job reads; each costs a fixed number of credits per call.
| Platform | Interface | What it returns | Price key | Credits |
|---|---|---|---|---|
| YouTube | Comments | The comments under one video: text, likes and reply count. | youtube.comments | 2 |
| YouTube | Search | Find videos by keyword: text, author and views. | youtube.search | 2 |
| TikTok | Shop product reviews | Get the first 10 buyers' reviews of a TikTok Shop product in the United States by the product's ID. | tiktok.product_reviews | 2 |
| TikTok | Shop product search | Search the products of TikTok Shop in the United States by terms. | tiktok.product_search | 2 |
| Xiaohongshu | Product reviews | Get the first page of buyers' reviews of a product of Xiaohongshu's shop by the product's ID. | xiaohongshu.product_reviews | 12 |
| Xiaohongshu | Product search | Search the products of Xiaohongshu's shop by terms. | xiaohongshu.product_search | 12 |
A worked example
A real run on the live service, recorded 2026-10-18: the question, the calls the AI chose and what it answered. The credits are what the run was charged.
What do viewers complain about in the comments of new Kindle reviews on YouTube?
| Platform | Tool | Sent | Requests | Returned | Credits |
|---|---|---|---|---|---|
| YouTube | comment_insights | targets: ["https://www.youtube.com/watch?v=f13Y9H7zooU","https://www.youtube.com/watch?v=GW-FpfCx0Js"] · focus: "what people think of the Kindle" · scale: "small" | 1 | 184 comments | 50 |
| Total | 1 | 184 items | 50 | ||
I read 184 comments under 2 review videos; 77 of the 181 that could be judged complain about something. The complaints that come up most:
- Prices that feel like a rip-off13%“this new one feels like a robbery.”24 comments
- No warm light10%“No warm light is a crime”19 comments
- Amazon's business practices7%“they locked their ecosystem and stole my books”12 comments
The posts it rests on, and the same calls as API requests, are in the whole recorded run.
50 credits is $0.17 at 300 credits per dollar (purchases from $10) and $0.10 at 500 per dollar (from $200). Failed calls are free, and a search that runs and finds nothing is charged as a call.
Give this to your AI
Connect SocialToAI once: in Claude, ChatGPT, Claude Code, Cursor and Codex you sign in and press Allow, with no key to paste; other clients are on the connect page. Then paste this and replace the parts in angle brackets. Never paste a key into a chat.
Use the SocialToAI tools (https://mcp.socialtoai.com/mcp) to mine reviews of <product>. Find recent review videos on YouTube and read their comments with the Comment analysis tool, then read the TikTok Shop reviews if it is sold there. Tell me what buyers praise and complain about, how often each comes up, and quote two buyers per point with links. End with the credits spent.Last updated: