Search result models: Video, Channel, and Playlist¶

Reference for the Video, Channel, and Playlist frozen dataclasses and the SearchResults paginated iterator returned by YouTube.search().

When you call YouTube.search() it returns a SearchResults object. Iterating over it yields one of three frozen dataclass models — Video, Channel, or Playlist — depending on what YouTube returned for that position in the results. All three models are immutable (frozen=True) and use __slots__ for memory efficiency.

Video¶

Represents a single video result from a search query.

video_id (str): The unique YouTube video identifier (e.g. "dQw4w9WgXcQ"). Never None.
title (str | None): The video title as displayed in search results. May be None if YouTube did not include it in the response.
channel (str | None): Display name of the uploading channel (e.g. "Rick Astley").
channel_id (str | None): The internal channel identifier that begins with UC (e.g. "UCuAXFkgsw1L7xaCfnd5JJOw"). Useful for passing to YouTube.channel().
duration (str | None): Human-formatted duration string as shown on YouTube (e.g. "10:23" or "1:02:47"). None for live streams or when unavailable.
views (int | None): View count as an integer (e.g. 1200000), parsed from the rendered text. The original wording is in views_text (str | None).
published_text (str | None): Relative publication date as shown in search results (e.g. "3 days ago", "2 years ago"). None for live streams. (Renamed from published in 2.0.)
published_at (datetime | None): Approximate UTC datetime computed from English published_text.
thumbnails (tuple[Thumbnail, ...]): All available sizes as Thumbnail(url, width, height).
thumbnail (str | None): URL of the highest-resolution thumbnail available in the search response.
description (str | None): Short description snippet from videoRenderer.descriptionSnippet.
url (str): Property. The canonical watch URL constructed from video_id: https://www.youtube.com/watch?v={video_id}.

Channel¶

Represents a single channel result from a search query.

channel_id (str): The internal channel identifier (e.g. "UCuAXFkgsw1L7xaCfnd5JJOw"). Never None.
title (str | None): The channel name (e.g. "Rick Astley").
handle (str | None): The channel's @handle if present in the search result (e.g. "@RickAstleyYT").
subscribers (int | None): Subscriber count as an integer. The original wording is in subscribers_text.
video_count (int | None): Formatted video count as YouTube renders it (e.g. "142 videos").
thumbnails (tuple[Thumbnail, ...]): All available sizes as Thumbnail(url, width, height).
thumbnail (str | None): URL of the channel's avatar thumbnail as returned in the search result.
url (str): Property. The canonical channel URL: https://www.youtube.com/channel/{channel_id}.

Playlist¶

Represents a single playlist result from a search query.

playlist_id (str): The unique playlist identifier (e.g. "PLbpi6ZahtOH6Ar_3GPy3workFpaikFpY3"). Never None.
title (str | None): The playlist title as shown in search results.
channel (str | None): Display name of the channel that owns the playlist.
video_count (int | None): The number of videos in the playlist as a string (e.g. "42").
thumbnails (tuple[Thumbnail, ...]): All available sizes as Thumbnail(url, width, height).
thumbnail (str | None): URL of the playlist's cover thumbnail.
url (str): Property. The canonical playlist URL: https://www.youtube.com/playlist?list={playlist_id}.

SearchResults¶

SearchResults is the lazy, paginated container returned by YouTube.search(). It transparently fetches new pages from YouTube as you consume items, so a simple for loop is all you need.

from ytscrape import YouTube

yt = YouTube()
results = yt.search("lo-fi music")

for item in results:
    print(type(item).__name__, item.title)

Properties¶

has_more (bool): True when YouTube has at least one more page of results that can be fetched. Becomes False once the continuation token is exhausted.

Methods¶

to_dict() / to_json() / to_csv()

Serialize every item the iterator would yield (max_results still applies). to_json() returns a JSON array; to_csv() a headered CSV table.

dump_json(path) / dump_csv(path)

Same payloads, written to a file path or file object.

fetch_next_page() (list[Video | Channel | Playlist])

Explicitly fetches the next page of results, appends them to the internal buffer, and returns the newly added items as a list. Returns an empty list when there are no more pages. Use this when you want fine-grained control over network calls instead of relying on the implicit iterator.

results = yt.search("guitar lessons")

# Fetch page 2 manually
page2 = results.fetch_next_page()
for item in page2:
    print(item.title)

Iteration¶

SearchResults is directly iterable. Each iteration step yields a Video, Channel, or Playlist instance. New pages are fetched automatically whenever the internal buffer runs out.

from ytscrape import YouTube
from ytscrape.models import Video, Channel, Playlist

yt = YouTube()

for item in yt.search("python tutorial"):
    if isinstance(item, Video):
        print(f"Video : {item.title}  [{item.duration}]  {item.url}")
    elif isinstance(item, Channel):
        print(f"Channel: {item.title}  ({item.subscribers_text})")
    elif isinstance(item, Playlist):
        print(f"Playlist: {item.title}  — {item.video_count} videos")

Limiting results¶

Pass max_results to YouTube.search() to cap how many items are yielded. The iterator stops once that number is reached, even if more pages exist.

# Collect at most 20 results without fetching unnecessary pages
top20 = list(yt.search("synthwave", max_results=20))

Tip

max_results limits the number of items yielded by the iterator, not the number of items fetched per page. YouTube returns roughly 20 results per page, so a max_results of 10 still fetches the first full page but stops yielding after 10 items.

AsyncSearchResults¶

Returned by await AsyncYouTube.search(). Same item types (Video, Channel, Playlist) and has_more / max_results behaviour as SearchResults, but iteration and paging are async:

from ytscrape import AsyncYouTube

async with AsyncYouTube() as yt:
    results = await yt.search("lo-fi music", max_results=20)
    async for item in results:
        print(item.title)
    page = await results.fetch_next_page()
    await results.dump_csv("search.csv")
    print(await results.to_json())

See the Async API guide.