YouTube Comment Replies Scraper API
Last updated: · REST · JSON
Send GET https://proxy.scrapier.io/api/v2/youtube-comment-replies/?continuation_tokens= with your key in the X-API-Key header. You get the thread back as JSON. 1 credit per thread; failed calls are free.
- 1 credit per thread
- Failed calls free
- 3 parallel requests per key
Call the YouTube Comment Replies Scraper API
Use the YouTube Comment Replies Scraper API from your own apps, scripts or agents. You need a Scrapier account and the API key from your dashboard. Pick a language.
import requests# Replace YOUR_API_KEY with the key from your Scrapier dashboardAPI_KEY = "YOUR_API_KEY"params = { "continuation_tokens": "", "max_items": "0", "language": "en", "region": "US",}r = requests.get( "https://proxy.scrapier.io/api/v2/youtube-comment-replies/", headers={"X-API-Key": API_KEY}, params=params, timeout=210,)r.raise_for_status()# The result sits under data.threadresult = r.json()["data"]["thread"]print(result)
// Replace YOUR_API_KEY with the key from your Scrapier dashboardconst API_KEY = "YOUR_API_KEY";const params = new URLSearchParams({ continuation_tokens: "", max_items: "0", language: "en", region: "US",});const res = await fetch(`https://proxy.scrapier.io/api/v2/youtube-comment-replies/?${params}`, { headers: { "X-API-Key": API_KEY }, signal: AbortSignal.timeout(210_000),});if (!res.ok) throw new Error(`HTTP ${res.status}`);const { data } = await res.json();console.log(data.thread);
# One GET request. The key goes in a header, never in the URL.GET /api/v2/youtube-comment-replies/?continuation_tokens=&max_items=0&language=en®ion=US HTTP/1.1Host: proxy.scrapier.ioX-API-Key: YOUR_API_KEYAccept: application/json
# Scrapier publishes an OpenAPI spec that covers every endpoint. No API key needed.curl -O https://proxy.scrapier.io/api/v2/openapi.json# Import openapi.json into Postman, Insomnia, Bruno or a code generator,# then add your key as the X-API-Key header.
# 1. Add the Scrapier MCP server to Claude Codeclaude mcp add -t http scrapier https://mcp.scrapier.io/mcp --header "X-API-Key: YOUR_API_KEY"# 2. Ask your agent. It finds the endpoint with search_endpoints# and calls it with run_scrape:"Get all replies for this YouTube comment thread token"run_scrape("youtube-comment-replies", {continuation_tokens: ""})
YouTube Comment Replies Scraper API in Python
Plain requests is all you need — no SDK. Every response is typed JSON.
Install
YouTube Comment Replies Scraper API in JavaScript
Works in Node 18+, Deno, Bun and edge runtimes with the built-in fetch.
Check your setup
YouTube Comment Replies Scraper API over HTTP
One GET request. Auth is a single X-API-Key header.
Check your setup
YouTube Comment Replies Scraper API OpenAPI definition
Download the spec once, then import it into the tools you already use. No key is needed to fetch it.
Download the spec
YouTube Comment Replies Scraper API over MCP
The same endpoint is a tool call for Claude Code, Cursor and Windsurf. Same continuation_tokens, same 1 credit per thread.
Install
Other ways to call it
YouTube Comment Replies Scraper API Interactive Playground
Edit the request, watch the code update, then run it. Without a key it shows the catalog sample and bills nothing. With your key it makes a real call from your browser; the key goes only to proxy.scrapier.io and is never stored.
GET /api/v2/youtube-comment-replies/?continuation_tokens=Press Run request to send it. Without a key you get the catalog sample.
Failed calls are not billed.
The sample is the same for every input: it shows the response shape, not the account you typed.
Example response, catalog sample from 9 Nov 2026
{
"success": true,
"data": {
"continuation_tokens": "Eg0SC3Y5UXRNNnFuRzUwGAYygwEaUBIaVWd4SHlJbmw3SXluZ1d2UjhnTjRBYUFCQWci...",
"thread": {
"input_continuation_token": "Eg0SC3Y5UXRNNnFuRzUwGAYygwEaUBIaVWd4SHlJbmw3SXluZ1d2UjhnTjRBYUFCQWci...",
"video_id": "v9QtM6qnG50",
"channel_id": "UCX6OQ3DkcsbYNE6H8uQQuVA",
"parent_comment_id": "UgxHyInl7IyngWvR8gN4AaABAg",
"video_title": "I Built A City To Save Kids From Illegal Labor",
"success": true,
"error": null,
"comments": [
{
"id": "UgxHyInl7IyngWvR8gN4AaABAg.AawtXymnHK1AawtnXUJNIO",
"content": "excellent job.",
"published_time_text": "3 days ago",
"published_time": "2026-09-20T08:08:34.750Z",
"reply_level": 1,
"author": {
"name": "@flow1n26",
"channel_id": "UCINudLBlbZCxSEugSRR7dCw",
"is_verified": false,
"is_creator": false,
"avatar_url": "https://yt3.ggpht.com/ytc/AIdro_lHq7nyOQOfims3HaL7WVdERI9RRMCkFIh5R21icJiVxc15lI8jurnAjKeeUVvfLfAcZw=s88-c-k-c0x00ffffff-no-rj",
"channel_url": "https://youtube.com/@flow1n26"
},
"engagement": {
"likes": 524,
"replies": 9
}
}
],
"reply_count": 1,
"next_continuation_token": null,
"has_more": false,
"scraped_at": "2026-09-23T08:08:42.049Z"
},
"next_cursor": null
}
}Data You Can Extract
YouTube Comment Replies Scraper extracts one reply thread per call: the video and parent comment it belongs to, a reply count, and every reply nested inside with its ID, text, time, reply level, author (name, channel ID, verified and creator flags, avatar) and engagement (likes and replies). Every field below comes from the catalog sample response, as typed JSON.
Thread identity
5 fields in the response.
- input_continuation_token
- video_id
- channel_id
- parent_comment_id
- video_title
Replies
2 fields in the response.
- comments
- reply_count
Status and resume
4 fields in the response.
- success
- error
- has_more
- next_continuation_token
Timestamp
1 field in the response.
- scraped_at
Response envelope and paging
3 fields in the response.
- success
- data.continuation_tokens
- data.next_cursor
Response Schema Reference
Every field in the sample response. Field names stay the same on every call.
| Field | Type | Example | Description |
|---|---|---|---|
thread.input_continuation_token | string | "Eg0SC3Y5UXRNNnFuRzUwGAYygwEaUBIaVWd4SHlJbmw…" | The token you sent |
thread.video_id | string | "v9QtM6qnG50" | ID of the video the thread is on |
thread.channel_id | string | "UCX6OQ3DkcsbYNE6H8uQQuVA" | UC channel ID of the video's channel |
thread.parent_comment_id | string | "UgxHyInl7IyngWvR8gN4AaABAg" | ID of the comment the replies belong to |
thread.video_title | string | "I Built A City To Save Kids From Illegal La…" | Title of the video |
thread.comments | array | 1 items | List of replies with id, content, published_time_text, published_time, reply_level, author and engagement |
thread.reply_count | integer | 1 | Number of replies returned |
thread.success | boolean | true | true when the thread was read |
thread.error | null | null | Error text when the thread failed; null otherwise |
thread.has_more | boolean | false | true when the thread was cut short |
thread.next_continuation_token | null | null | Token to resubmit when the thread was truncated; null when complete |
thread.scraped_at | string | "2026-09-23T08:08:42.049Z" | When Scrapier read the thread (ISO 8601, UTC) |
success | boolean | true | true on every 200 response |
data.continuation_tokens | string | "Eg0SC3Y5UXRNNnFuRzUwGAYygwEaUBIaVWd4SHlJbmw…" | Your continuation token, echoed back |
data.next_cursor | null | null | Always null here: a thread is a single lookup |
Request Parameters Reference
| Name | Type | Required | Default | Description | Example |
|---|---|---|---|---|---|
continuation_tokens | string | Yes | — | Base64url reply continuation token from a "Show more replies" call. One per call. Required. Aliases: continuation_token, token, tokens | |
cursor | string | No | — | Pass back data.next_cursor from the previous response to get the next page; null means the last page | — |
max_items | integer | No | 0 | Replies per thread. 0 or omitted walks the thread to the end. Locked in the cursor | 0 |
language | string | No | en | YouTube interface language (hl), default en. Locked in the cursor | en |
region | string | No | US | YouTube region (gl), default US. Locked in the cursor | US |
use_proxy | boolean | No | true | Route through Scrapier's residential proxy pool. Leave as is | true |
raw_data | boolean | No | false | Return YouTube's raw upstream payload under raw_responses instead of the formatted items | true |
One token per call: to expand ten threads, make ten calls. Booleans are the strings true or false.
Pagination in the YouTube Comment Replies Scraper API
The YouTube Comment Replies Scraper API returns results in pages. Pass data.next_cursor back as cursor to get the next page; when it is null there are no more. Each successful call, first page or later, costs 1 credit.
Access Without YouTube Login
The YouTube Comment Replies Scraper API needs no YouTube account, no login and no session cookies. You send your Scrapier key in the X-API-Key header and get JSON back. It reads public data only: what any visitor sees without logging in.
Speed, Cost and Concurrency
| Metric | Value | Source |
|---|---|---|
| Cost of a successful call | 1 credit | Scrapier catalog |
| Cost of a failed call | 0 credits | Any non-2xx response |
| Parallel requests per key | 3 | Above that: 429 with Retry-After |
| Recommended client timeout | 210 seconds | Scrapier docs, every endpoint |
Latency of the YouTube Comment Replies Scraper API
Scrapier does not publish a latency or success-rate figure for this endpoint. Each call reads YouTube at call time, so response time depends on YouTube and on the size of the result. Set a client timeout of 210 seconds, as the docs recommend for every endpoint.
Limits of the Endpoint
Honest constraints of the YouTube Comment Replies Scraper API.
- Public data only.
- A failed call, for example a missing or unavailable thread, returns a non-2xx status and costs 0 credits.
- Data is read live at call time.
- 3 parallel requests per key; above that the API answers
429withRetry-After. - Large results come in pages; pass
data.next_cursorback ascursor. - Fields YouTube does not show are not returned.
Scrapier vs the Official YouTube API
The official YouTube Data API is the better choice when you reply to or moderate comments on your own channel, which a read-only scraper cannot do.
| Feature | Scrapier thread scraper | Official YouTube API |
|---|---|---|
| Access needed | One Scrapier API key | A Google Cloud project, an API key and a daily quota |
| Which replies | Every reply in a thread, from a continuation token | Replies of a parent comment through the comments resource, in pages |
| Fields returned | 12 thread fields with replies, authors and likes nested | Depends on the resource and parts you request |
| Rate limits | Up to 3 parallel requests per key; 429 with Retry-After | Daily quota units set per Google Cloud project |
| Setup | Sign up, copy a key, send a request | Create a project, enable the API, create credentials |
| Cost | 1 credit per thread, from $1.20 per 1,000 threads | Free within quota; more needs a quota extension request |
Errors and Retry Handling
| Status | Meaning | Billed | Fix |
|---|---|---|---|
| 200 | Thread data returned in data.thread | Yes | — |
| 400 | Bad or missing input | No | Check continuation_tokens: one value per call |
| 401 | X-API-Key header missing | No | Add the header |
| 402 | Out of credits (body has balance) | No | Top up; retrying will not help |
| 403 | Invalid or revoked API key | No | Check the key in your dashboard |
| 429 | Throttled | No | Wait Retry-After seconds, then retry |
| 5xx | Upstream failure or timeout | No | Retry with backoff; set a 210 s client timeout |
From the error reference. Any non-2xx response costs 0 credits.
Retry logic for the YouTube Comment Replies Scraper API
import time, requests def get_comment_replies(value, key, tries=5): for attempt in range(tries): r = requests.get( "https://proxy.scrapier.io/api/v2/youtube-comment-replies/", headers={"X-API-Key": key}, params={"continuation_tokens": value}, timeout=210, ) if r.status_code == 429: time.sleep(int(r.headers.get("Retry-After", 2 ** attempt))) elif r.status_code >= 500: time.sleep(2 ** attempt) else: r.raise_for_status() return r.json() raise RuntimeError("gave up after retries")
Pricing and Rate Limits
| Plan | Monthly price | Threads included | Cost per 1,000 |
|---|---|---|---|
| Hobby | $29 | 15,000 | $1.93 |
| Pro | $79 | 50,000 | $1.58 |
| Scale | $299 | 250,000 | $1.20 |
Billing rules for the YouTube Comment Replies Scraper API
The YouTube Comment Replies Scraper API bills 1 credit per successful thread. Failed calls, meaning any non-2xx status, cost 0 credits. Every plan allows 3 parallel requests per key. The free tier lets you try it before you pay.
Bulk Lookups and CSV
continuation_tokens takes one value per call. For a list, loop and run up to 3 requests in parallel per key, then flatten each row into one CSV line.
import csv, requests from concurrent.futures import ThreadPoolExecutor values = [""] # your list def fetch(value): r = requests.get( "https://proxy.scrapier.io/api/v2/youtube-comment-replies/", headers={"X-API-Key": "YOUR_API_KEY"}, params={"continuation_tokens": value}, timeout=210, ) r.raise_for_status() out = r.json()["data"]["thread"] return out if isinstance(out, list) else [out] # 3 parallel requests per key is the limit with ThreadPoolExecutor(max_workers=3) as pool: rows = [row for part in pool.map(fetch, values) for row in part] cols = ["video_id", "video_title", "parent_comment_id", "reply_count", "has_more", "success"] with open("out.csv", "w", newline="") as f: w = csv.DictWriter(f, fieldnames=cols, extrasaction="ignore") w.writeheader() w.writerows(rows)
CSV export from the YouTube Comment Replies Scraper API
video_id,video_title,parent_comment_id,reply_count,has_more,success v9QtM6qnG50,I Built A City To Save Kids From Illegal Labor,UgxHyInl7IyngWvR8gN4AaABAg,1,false,true
AI Agent Access via MCP
MCP for the YouTube Comment Replies Scraper API
The same endpoint is available to AI agents through the Scrapier MCP server at https://mcp.scrapier.io/mcp. One command registers it in Claude Code: claude mcp add -t http scrapier https://mcp.scrapier.io/mcp --header "X-API-Key: YOUR_API_KEY". The agent finds the endpoint with search_endpoints and calls it with run_scrape and the endpoint id youtube-comment-replies. It takes the same continuation_tokens input and costs the same 1 credit per thread. Setup for every client is on the YouTube Comment Replies Scraper MCP page.
Integration Paths by Role
n8n and Make
- Add an HTTP Request step
- GET the endpoint with your
X-API-Keyheader - Map the
data.threadfields to your columns
Loop and paginate
MCP tool in your agent
Call run_scrape with endpoint youtube-comment-replies from Claude, Cursor or Windsurf. Setup is on the MCP page.
CSV and flattened JSON
Flatten each row into one CSV line, one column per field, ready for Sheets, Excel or a BI tool.
Legal and Compliance Notes
Scraping publicly available YouTube data is generally lawful in many countries, but the answer depends on what you collect, where you are and how you use it. In the US, hiQ Labs v. LinkedIn is the case most often cited on public-data scraping. Scrapier returns only data any visitor can see without logging in. Under GDPR and similar laws, collecting public data can be permissible, but how you store and use personal data, such as names or comment text, must still comply with local data-protection rules, and YouTube's terms apply. Consult legal counsel if your use case involves bulk storage of personal data. This is not legal advice; see our Terms of Use.
Frequently Asked Questions
Is there a YouTube thread API?
Yes. Scrapier's YouTube Comment Replies Scraper API returns any public thread as JSON with one GET request and an X-API-Key header. No YouTube app or login.
How do I call it from Python?
Use <code>requests.get</code> on <code>https://proxy.scrapier.io/api/v2/youtube-comment-replies/</code> with <code>params={'continuation_tokens': ''}</code> and <code>headers={"X-API-Key": KEY}</code>. Read <code>r.json()["data"]["thread"]</code>.
What happens when a call fails?
Any non-200 response is free. Retry 429 and 5xx with backoff; the API already retries upstream failures.
How much does it cost?
1 credit per successful call — about $1.58 per 1,000 threads on the Pro plan. Failed calls are free.
Is there an OpenAPI spec?
Yes. It needs no key and covers every Scrapier endpoint.
Related YouTube APIs
Comments API
Comments on a video, top or newest first, with author, likes, reply count and creator heart.
Video Details API
One video by URL: title, channel, description, views, likes, duration, tags and upload date.
Channel Video API
A channel's videos, Shorts or both, a page at a time, with title, views, duration and publish time.
All YouTube APIs
All 17 YouTube endpoints.
Fetch public YouTube threads as typed JSON in one request
1 credit per thread · failed calls free