---
title: "How do I run cross-platform comment sentiment and pain-point analysis with AI and MCP? | FeedMCP"
description: "Connect FeedMCP to Claude or Cursor to pull normalized comment objects across TikTok ($0.002), Douyin ($0.002), Reddit ($0.002), YouTube ($0.002) and Xiaoh"
canonical: "https://feedmcp.com/answers/social-media-sentiment-analysis-ai/"
---
<p class="crumbs"><a href="/answers/">Answers</a></p>

# How do I run cross-platform comment sentiment and pain-point analysis with AI and MCP?

Connect FeedMCP to Claude or Cursor to pull normalized `comment` objects across TikTok ($0.002), Douyin ($0.002), Reddit ($0.002), YouTube ($0.002) and Xiaohongshu ($0.02) in one prompt, then cluster sentiment, complaints and feature requests automatically.

Because every platform uses the exact same comment schema, one prompt or script covers both global and Chinese channels for Voice-of-Customer (VOC) analysis. See our [comment mining guide](/guides/social-comment-pain-point-mining/).

## Endpoints and pricing

| Endpoint | ID | Price | What it returns |
| --- | --- | --- | --- |
| [TikTok video comments](/apis/tiktok/videos/comments/list/) | `tiktok.videos.comments.list` | $0.002 | Comments under a video: text, time, commenter, likes and reply count, page by page |
| [Xiaohongshu note comments](/apis/xiaohongshu/notes/comments/list/) | `xiaohongshu.notes.comments.list` | $0.02 | Comments under a note: text, time, commenter, likes and reply count, page by page |
| [Douyin video comments](/apis/douyin/videos/comments/list/) | `douyin.videos.comments.list` | $0.002 | Comments under a video: text, time, commenter, likes and reply count, page by page |
| [Reddit post comments](/apis/reddit/posts/comments/list/) | `reddit.posts.comments.list` | $0.002 | Top-level comments under a post: text, time, author and score, page by page |
| [YouTube video comments](/apis/youtube/videos/comments/list/) | `youtube.videos.comments.list` | $0.002 | Comments under a video: text, time, commenter, likes and replies, top or newest first |

Pay per call with failed calls free. Verify live prices for free via `GET https://api.feedmcp.com/v1/catalog/prices` and cap per-call spend with `X-Max-Price`.

## How to do it

### Option 1: Ask your AI via MCP

1. Sign up in the [console](/console/keys/) ($0.10 free trial credit, no card) and create an API key with a spend cap.
2. Add `https://api.feedmcp.com/mcp` to Claude, Cursor, Codex or Cherry Studio via the [MCP setup page](/mcp/) or [Integrations](/integrations/).
3. Ask your AI:

> Pull comments from these TikTok, Xiaohongshu and Reddit posts and compare complaints between global and Chinese users

### Option 2: Call the HTTP API directly

```bash
curl "https://api.feedmcp.com/v1/tiktok/videos/comments/list" \
  -H "Authorization: Bearer $FEEDMCP_API_KEY" \
  -H "X-Max-Price: 0.002"
```

## Key facts at a glance

| Item | Detail |
| --- | --- |
| Starting price | $0.002 and up (pay per call, no subscription) |
| Failed calls | Free (`charged: false`) |
| Spend guardrails | Per-key spend cap + `X-Max-Price` + `Idempotency-Key` |
| Data scope | Public data only; no platform account login required |
| Access | Remote MCP (`Streamable HTTP`) + REST API + OpenAPI |

## Questions

### How do I run cross-platform comment sentiment and pain-point analysis with AI and MCP?

Connect FeedMCP to Claude or Cursor to pull normalized `comment` objects across TikTok ($0.002), Douyin ($0.002), Reddit ($0.002), YouTube ($0.002) and Xiaohongshu ($0.02) in one prompt, then cluster sentiment, complaints and feature requests automatically.

### How much do these endpoints cost? Are failed calls billed?

Endpoints on this page start at $0.002 per call with no monthly subscription and non-expiring balances. Bad parameters, timeouts, upstream errors and missing objects are free. New accounts receive $0.10 trial credit, no card required.

### Do I need my own platform account or cookies?

No. FeedMCP reads public data only; calls never go through your personal accounts.

## Related answers

- [How do FeedMCP, DataXAPI, Qovia and AIsa compare for social media data and MCP?](/answers/feedmcp-vs-dataxapi-vs-qovia-vs-aisa/)
- [How to use TikHub-style Douyin, Xiaohongshu and TikTok endpoints in Claude or Cursor via a hosted MCP server?](/answers/tikhub-mcp-server-and-unified-api/)
- [How to connect OpenClaw, Antigravity, Windsurf or Grok agents to a social media data MCP server?](/answers/connect-openclaw-antigravity-grok-to-social-media-mcp/)
- [How to extract video transcripts and subtitles from TikTok, YouTube, Instagram, X, Reddit and Meta Ads via API or MCP?](/answers/extract-video-transcripts-subtitles-api-mcp/)
- [Can ChatGPT read Xiaohongshu (RedNote) notes and comments?](/answers/can-chatgpt-read-xiaohongshu/)
- [How do I let Claude or Cursor read Xiaohongshu notes, comments and PGY data?](/answers/can-claude-read-xiaohongshu/)

---

HTML version: https://feedmcp.com/answers/social-media-sentiment-analysis-ai/
