# YouTube Comments MCP Server — Claude & Cursor | Scrapier (https://scrapier.io/scrapers/youtube/comments/mcp/)

> Scrapier's YouTube Comments Scraper MCP server gives Claude, Cursor and any MCP client live video comments as JSON. One command connects it. 1 credit per page.

Last updated: 2026-11-10

**GET**/api/v2/youtube-comments-scraper/

## YouTube Comments Scraper _MCP Server_

Give Claude, Cursor and ChatGPT a tool for any public YouTube page of comments. One command connects it; every call returns typed JSON. 1 credit per page of comments.

[Get a free API key →](https://app.scrapier.io/sign-up) [Jump to setup](#setup)

Free API key · 1 credit per page of comments · failed calls free

Last updated 9 Nov 2026

Live agent run

Copy for your AI agent

Install

## Connect in one command

Quick answerScrapier's MCP server gives **Claude, Cursor and any MCP client** tools for public YouTube pages of comments, returned as JSON. One command connects it. Each call costs 1 credit.

terminal

claude mcp add -t http scrapier \\
  https://mcp.scrapier.io/mcp \\
  --header "X-API-Key: YOUR\_API\_KEY"

Then ask: _"Get the YouTube page of comments for https://www.youtube.com/watch?v=uKv3mi2t2KE"_

Connect

## Custom connector

Connect any client that supports remote MCP with one URL and your API key.

### Custom connector

Any MCP client

Paste one URL into any client that supports remote MCP and send your Scrapier API key in an `X-API-Key` header.

`https://mcp.scrapier.io/mcp`

Where to paste it

- **Claude**Settings → Connectors → Add
- **ChatGPT**Settings → Apps → Developer mode → Create
- **Cursor**Settings → MCP → Add new server

[Step-by-step setup ↓](#setup)

Setup

## Setup in Claude, Cursor, Windsurf and ChatGPT

Pick your client. Three steps, under a minute.

1. Get a free API key from your Scrapier dashboard.
2. Run the command below in your terminal.
3. Restart Claude Code and run `/mcp` to see the tools.

terminal

claude mcp add -t http scrapier https://mcp.scrapier.io/mcp --header "X-API-Key: YOUR\_API\_KEY"

claude · ~/project

**›** /mcp

Manage MCP servers

● scrapier · connected · 6 tools

└ search\_endpoints

└ run\_scrape

└ …4 more

**›**

### Verify the connection

Ask **"List your YouTube tools"**. The agent should find `run_scrape` and `search_endpoints`.

Tools

## YouTube MCP Tools

Use `search_endpoints` to discover, `run_scrape` to call.

| Tool | Returns | Inputs | Key fields | Credits |
| --- | --- | --- | --- | --- |
| `run_scrape` | Run the page of comments scraper | `endpoint="youtube-comments-scraper", urls="..."` | cid, type, published\_time\_text, comment, author | 1 |

### Output fields

run\_scrape · 200 OK

{
  "success": true,
  "data": {
    "urls": "https://www.youtube.com/watch?v=uKv3mi2t2KE",
    "video\_id": "uKv3mi2t2KE",
    "page": 1,
    "page\_size": 20,
    "comments": \[
      {
        "cid": "Ugx3cSNbOeJMcSWom1R4AaABAg",
        "type": "comment",
        "published\_time\_text": "7 months ago",
        "comment": "🚨 Watch our three part episode series on the Student Revolution that took Bangla",
        "author": "@thedeshbhakt",
        "author\_channel\_id": "UCmTM\_hPCeckqN3cPWtYZZcg",
        "avatar": "https://yt3.ggpht.com/DfVqzaxetTmlyiL2oZHng3i9nKBEKJWWjQoPVUYf3U9Qq6rxDeCtaXjCUj9ARjqdluzwbv\_r=s88-c-k-c0x00ffffff-no-rj",
        "author\_is\_channel\_owner": true,
        "is\_verified": true,
        "vote\_count": 264,
        "has\_creator\_heart": true,
        "reply\_count": 72,
        "reply\_to\_cid": null,
        "video\_id": "uKv3mi2t2KE",
        "page\_url": "https://www.youtube.com/watch?v=uKv3mi2t2KE",
        "title": "BNP Wins Big In Historic Elections | Options For India In Post-Hasina Bangladesh? | Akash Banerjee",
        "comments\_count": 8300,
        "comment\_id": "Ugx3cSNbOeJMcSWom1R4AaABAg",
        "like\_count": 264,
        "input\_url": "uKv3mi2t2KE",
        "scraped\_at": "2026-09-14T08:58:22.312Z"
      }
    \],
    "next\_cursor": "eyJ2IjoidUt2M21pMnQyS0UiLCJ0IjoiRWtNU0MuLi4iLCJwIjoxfQ"
  }
}

Examples

## Example Prompts and Tool Calls

The prompt, the tool call, the JSON and the answer.

### Get page of comments data

Get the page of comments data for https://www.youtube.com/watch?v=uKv3mi2t2KE.

### ⚙ run\_scrape(endpoint="youtube-comments-scraper", urls="https://www.youtube.com/watch?v=uKv3mi2t2KE")

run\_scrape(endpoint="youtube-comments-scraper", urls="https://www.youtube.com/watch?v=uKv3mi2t2KE")

{
  "cid": "Ugx3cSNbOeJMcSWom1R4AaABAg",
  "type": "comment",
  "published\_time\_text": "7 months ago",
  "comment": "🚨 Watch our three part episode series on the Student Revolution that took Bangla"
}

Here are the page of comments details from the Scrapier API.

Workflows

## Agent Workflows

### YouTube data pipeline

1. input list
2. run\_scrape
3. filter results
4. export

Feed a list, pull each page of comments, filter by criteria, export the result.

Compare

## YouTube Scraper MCP vs the Official YouTube MCP

|  | Scrapier YouTube MCPThis page | Official YouTube API | Apify actor over MCP |
| --- | --- | --- | --- |
| Data returned | Any public page of comments, typed JSON | Your own accounts and insights | Varies by actor |
| Auth needed | One Scrapier key | YouTube app + business account | Apify token |
| Cost model | 1 credit per call | Free once approved | Compute units per run |
| Rate limits | 3 parallel per key | Per-user hourly quota | Per actor |
| Best for | Research, enrichment, monitoring | Own-account analytics | Custom one-off crawls |

### Limits

- Public data only — private pages of comments return identity fields only
- Counts are live at call time; no history unless you store it

Billing

## Credits, Rate Limits and Security

Credits per call**1**

Parallel requests**3 per key**

Auth**One Scrapier key, no YouTube login**

- Your key stays in your client config, never in prompts
- No YouTube account or cookies are used
- Failed calls cost 0 credits
- Rotate or revoke keys any time from the dashboard

[See full pricing →](https://scrapier.io/pricing/)

Troubleshooting

## Troubleshooting

### Tools not showing

Restart the client after editing its config. In Claude Code run /mcp; in Cursor check the server toggle is on.

### 401 invalid key

The key is missing or mistyped. Copy it again from the dashboard.

### 429 rate limit

You exceeded your plan's parallel requests. The agent should wait for Retry-After.

### Timeout on large requests

Batch in smaller groups, or raise the client timeout to 210 seconds.

FAQ

## Frequently Asked Questions

### Is there a YouTube scraper MCP server?

Yes. Scrapier's MCP server exposes YouTube page of comments tools to Claude, Cursor, Windsurf and ChatGPT. One command connects it and every call returns typed JSON.

### How do I scrape YouTube pages of comments with Claude?

Run \<code>claude mcp add -t http scrapier https://mcp.scrapier.io/mcp\</code> with your API key, then ask Claude for any public page of comments. It calls \<code>run\_scrape\</code> and reads the JSON directly.

### Does it need a YouTube login?

No. The server reads public data only and uses no YouTube account or cookies.

### How much does a YouTube MCP call cost?

1 credit per call — about $1.58 per 1,000 pages of comments on the Pro plan. Failed calls are free.

### Can ChatGPT use the YouTube scraper?

Yes. Add Scrapier as a custom connector in ChatGPT's developer mode and the same tools appear in chat.

More servers

## Related MCP Servers

[### Channel Info MCP

channel →](https://scrapier.io/scrapers/youtube/channel/mcp/)[### Channel Live/Streams MCP

channel-live →](https://scrapier.io/scrapers/youtube/channel-live/mcp/)[### Channel Video Scraper MCP

channel-videos →](https://scrapier.io/scrapers/youtube/channel-videos/mcp/)[### All YouTube scrapers

17 endpoints →](https://scrapier.io/scrapers/youtube/)

## Give your agent YouTube data

1 tool · JSON · one key

[Get a free API key →](https://app.scrapier.io/sign-up)[See pricing](https://scrapier.io/pricing/)

---
Canonical page: https://scrapier.io/scrapers/youtube/comments/mcp/
Site index for AI agents: https://scrapier.io/llms.txt
