YouTube via Scavio

Search YouTube and retrieve videos, shorts, search suggestions, video metadata, comments, transcripts, related videos, download streams, and full channel data (info, videos, shorts, community posts). All endpoints return structured JSON.

When to trigger

Use this skill when the user asks to:

• Find YouTube videos, shorts, or channels on a topic
• Get a video's metadata, view count, duration, or captions list
• Read a video's comments and comment replies
• Pull a full transcript or timed subtitles for a video
• Get related videos or direct download stream URLs
• Look up a channel's info, videos, shorts, or community posts
• Resolve a channel handle or URL to a channel ID

Setup

Get a free API key at scavio.dev (50 free credits to get started, no card required):

export SCAVIO_API_KEY=sk_live_your_key

Every request is a POST with a JSON body and:

Authorization: Bearer $SCAVIO_API_KEY

Endpoints

Base URL: https://api.scavio.dev. All paths are under /api/v1/youtube.

Video endpoints

Channel endpoints

POST /api/v1/youtube/metadata still works as a deprecated alias of /video. Use /video for new code.

Workflow

1. Find a video: call /search with search. Use sort_by: view_count for the most-watched result. Read results[].video_id.
2. Get metadata: call /video with video_id (accepts a full watch URL too) for description, length_seconds, view_count, keywords, and captions[].
3. Read discussion: call /comments, then /comments/replies with a comment's reply_cursor to expand a thread.
4. Get the text: call /transcript for a plain transcript (format: text) or timed subtitles (format: srt).
5. Download: call /streams for direct format URLs.
6. Channels: resolve a handle with /channel/resolve, then call /channel, /channel/videos, /channel/shorts, or /channel/community.

Paginated endpoints return next_cursor and has_more; pass next_cursor back as cursor for the next page.

Parameters

Search (/search)

Legacy boolean flags (subtitles, creative_commons, hd, 4k, live, 360, 3d, hdr, vr180) are still accepted for backward compatibility; prefer the features array.

Shorts (/shorts)

search* , sort_by, cursor.

Suggestions (/suggestions)

search* , language (default en), region (default US).

Video (/video) and metadata alias

video_id* — accepts a raw ID or a full watch URL.

Comments (/comments)

video_id* , cursor.

Comment replies (/comments/replies)

video_id* , reply_cursor* (a comment's reply_cursor), cursor.

Transcript (/transcript)

video_id* , language (default en), format (text for plain, srt for timed; default text).

Related (/related)

video_id* , cursor.

Streams (/streams)

video_id* .

Channel search (/channel/search)

search* , cursor.

Channel (/channel)

channel_id* — accepts a channel ID, @handle, or channel URL.

Channel videos / shorts / community

channel_id* , cursor.

Channel resolve (/channel/resolve)

channel* — a @handle or channel URL. Returns channel_id and channel_url.

Examples

import requests

BASE = "https://api.scavio.dev"
# Your key from https://scavio.dev. Load it from your environment or secret
# store in real code - keep it out of source control.
API_KEY = "sk_your_key_here"
HEADERS = {"Authorization": f"Bearer {API_KEY}"}

# 1. Search — use "search" field, not "query"
search = requests.post(f"{BASE}/api/v1/youtube/search", headers=HEADERS,
    json={"search": "langchain tutorial", "type": "video", "sort_by": "view_count"}).json()

video_id = search["data"]["results"][0]["video_id"]

# 2. Full video metadata (captions list included)
video = requests.post(f"{BASE}/api/v1/youtube/video", headers=HEADERS,
    json={"video_id": video_id}).json()

# 3. Transcript as plain text
transcript = requests.post(f"{BASE}/api/v1/youtube/transcript", headers=HEADERS,
    json={"video_id": video_id, "language": "en", "format": "text"}).json()

# 4. Resolve a handle, then pull the channel's videos
resolved = requests.post(f"{BASE}/api/v1/youtube/channel/resolve", headers=HEADERS,
    json={"channel": "@freecodecamp"}).json()
channel_id = resolved["data"]["channel_id"]

videos = requests.post(f"{BASE}/api/v1/youtube/channel/videos", headers=HEADERS,
    json={"channel_id": channel_id}).json()

Response shapes

Every response uses the envelope { data, response_time, credits_used, credits_remaining }. Key data fields per endpoint:

• search — results[] (type, video_id, title, url, description_snippet, thumbnail, duration_text, view_count, published_time, channel{id,name,url}), plus shorts, channels, playlists, next_cursor, has_more.
• shorts — results[] (video_id, title, url, thumbnail, view_count, published_time, author, channel_id), next_cursor, has_more.
• suggestions — suggestions[] (strings), total_count.
• video — video_id, title, author, channel_id, channel_url, published_at, description, length_seconds, view_count, keywords[], thumbnail, playability_status, chapters, captions[] (language_code, language_name, url).
• comments — comments[] (comment_id, text, like_count, reply_count, published_time, reply_cursor, author{channel_id,name,url,avatar,is_verified,is_creator}), next_cursor, has_more.
• comments/replies — replies[] (same item shape as a comment), next_cursor, has_more.
• transcript — video_id, language_code, language_name, format, content.
• related — results[] (video_id, title, url, author, channel_id, channel_url, thumbnail, view_count, published_time, length_seconds), total_count.
• channel/search — results[] (channel_id, name, handle, url, thumbnail, subscriber_count, description, verified), next_cursor, has_more, total_count.
• channel — channel_id, title, description, handle, url, subscriber_count, video_count, view_count, country, creation_date, verified, has_business_email, avatar, banner, links[].
• channel/videos — channel_id, results[] (video_id, title, url, thumbnail, duration_text, view_count, published_time, is_live), next_cursor, has_more.
• channel/shorts — channel_id, results[] (video_id, title, url, thumbnail, view_count), next_cursor, has_more, total_count.
• channel/community — channel_id, posts[] (post_id, url, text, author_name, author_channel_id, published_time, vote_count, comment_count, images[], attachment_type), next_cursor, has_more.
• channel/resolve — channel_id, channel_url.
• streams — video_id, title, author, length_seconds, view_count, is_live, formats[], adaptive_formats[], available_qualities[], expires_in_seconds.

{
  "data": {
    "results": [
      {
        "type": "video",
        "video_id": "sVcwVQRHIc8",
        "title": "Learn RAG From Scratch - Python AI Tutorial",
        "url": "https://www.youtube.com/watch?v=sVcwVQRHIc8",
        "duration_text": "2:33:11",
        "view_count": 1258310,
        "published_time": "1 year ago",
        "thumbnail": "https://i.ytimg.com/vi/sVcwVQRHIc8/hq720.jpg",
        "channel": { "id": "UC8butISFwT-Wl7EV0hUK0BQ", "name": "freeCodeCamp.org", "url": "https://www.youtube.com/@freecodecamp" }
      }
    ],
    "next_cursor": "EpMD...",
    "has_more": true
  },
  "credits_used": 2,
  "credits_remaining": 998
}

Guardrails

• The search parameter is search, not query — this is different from other Scavio endpoints.
• Credits are not uniform: search and shorts cost 2, streams costs 3, transcript costs 8, and every other endpoint costs 1. Inform the user before paginating through many pages, especially transcripts.
• /metadata is a deprecated alias of /video; use /video in new code.
• Stream URLs from /streams expire (expires_in_seconds) — use them promptly.
• Never fabricate video IDs, view counts, transcripts, comments, or channel data. Only return API data.

Failure handling

• 400 means an invalid parameter (e.g. a missing video_id or channel_id) — fix and retry.
• 401 means the API key is invalid or missing. Check SCAVIO_API_KEY.
• 429 means rate or usage limit exceeded. Wait before retrying. See https://scavio.dev/docs/rate-limits.
• 502 / 503 mean upstream is temporarily unavailable. Wait a few seconds before retrying.
• If search returns no results, suggest different keywords or relaxing filters.
• If a transcript is unavailable in the requested language, retry with language: en or check /video captions[] for available languages.
• If SCAVIO_API_KEY is not set, prompt the user to export it before continuing.

LangChain

pip install langchain-scavio==4.0.2

from langchain_scavio import ScavioSearchTool
tool = ScavioSearchTool(engine="youtube")