Inoue-AI YouTube SDK

Multi-language SDK for the YouTube Data API v3 (and Analytics API v2 in Python). Ships both a Python client (under youtube/) and a Go client (under go/).

Note: This repository was renamed from Inoue-AI/YouTube-Python-SDK to Inoue-AI/Inoue-AI-YouTube-SDK. GitHub auto-redirects the old URL, so existing pip install ... @ git+https://github.com/Inoue-AI/YouTube-Python-SDK.git commands continue to work. The local clone keeps the legacy directory name.

Python

The Python source remains under youtube/. Install:

pip install "git+https://github.com/Inoue-AI/Inoue-AI-YouTube-SDK.git"

Go

The Go SDK lives under go/. It exposes a focused subset matching the calls the Inoue AI Go backend makes today (channel + uploads playlist + single video). Phase 3 will expand the surface as needed.

go get github.com/Inoue-AI/Inoue-AI-YouTube-SDK/go@latest

client := youtube.New(youtube.ClientOptions{AccessToken: "..."})
defer client.Close()
ch, err := client.GetChannel(ctx, youtube.GetChannelParams{Mine: true})

See go/README.md for the full Go API.

---

Original Python SDK

Async Python SDK for the YouTube Data API v3 and YouTube Analytics API v2.

Built with aiohttp and pydantic v2. Follows the same architecture as the TikTok Python SDK.

Features

• 14 API namespaces — Videos, Channels, Playlists, PlaylistItems, Comments, CommentThreads, Subscriptions, Thumbnails, Search, Captions, ChannelSections, ChannelBanners, Watermarks, Analytics
• Pydantic v2 models for every request and response — full type safety, IDE autocomplete, and serialization
• Async-first — powered by aiohttp with lazy session creation
• Resumable video uploads — handles chunked uploads automatically
• Auto-pagination — async iterators for paging through large result sets
• Dual auth — OAuth 2.0 access tokens (full access) or API keys (public read-only)
• Comprehensive error hierarchy — typed exceptions mapped from YouTube error codes

Installation

pip install youtube-python-sdk

Or install from source:

git clone https://github.com/inoue-ai/youtube-python-sdk.git
cd youtube-python-sdk
pip install -e ".[dev]"

Quick Start

Read public data with an API key

import asyncio
from youtube import YouTubeClient, VideoPart

async def main():
    async with YouTubeClient(api_key="YOUR_API_KEY") as yt:
        response = await yt.videos.list(
            parts=[VideoPart.SNIPPET, VideoPart.STATISTICS],
            chart="mostPopular",
            max_results=10,
        )
        for video in response.items:
            print(f"{video.snippet.title} — {video.statistics.view_count} views")

asyncio.run(main())

Manage your channel with OAuth

import asyncio
from youtube import (
    YouTubeClient, ChannelPart, Video, VideoSnippet, VideoStatus,
    PrivacyStatus, VideoPart,
)

async def main():
    async with YouTubeClient(access_token="OAUTH_ACCESS_TOKEN") as yt:
        # Get your channel info
        channels = await yt.channels.list(
            parts=[ChannelPart.SNIPPET, ChannelPart.STATISTICS],
            mine=True,
        )
        channel = channels.items[0]
        print(f"Channel: {channel.snippet.title}")
        print(f"Subscribers: {channel.statistics.subscriber_count}")

        # Upload a video
        video = await yt.videos.insert(
            body=Video(
                snippet=VideoSnippet(
                    title="My New Video",
                    description="Uploaded via the YouTube Python SDK!",
                    tags=["sdk", "python", "youtube"],
                    category_id="22",
                ),
                status=VideoStatus(
                    privacy_status=PrivacyStatus.PRIVATE,
                ),
            ),
            file_path="/path/to/video.mp4",
        )
        print(f"Uploaded: {video.id}")

        # Set a custom thumbnail
        await yt.thumbnails.set_from_file(
            video_id=video.id,
            file_path="/path/to/thumbnail.jpg",
        )

asyncio.run(main())

Search

from youtube import YouTubeClient, SearchResultType, SearchOrder

async with YouTubeClient(api_key="KEY") as yt:
    results = await yt.search.list(
        q="python tutorial",
        type=SearchResultType.VIDEO,
        order=SearchOrder.VIEW_COUNT,
        max_results=25,
    )
    for item in results.items:
        print(f"{item.snippet.title} — {item.id.video_id}")

Playlists

from youtube import (
    YouTubeClient, Playlist, PlaylistSnippet, PlaylistStatus,
    PlaylistPart, PlaylistItem, PlaylistItemSnippet, PlaylistItemPart,
    PrivacyStatus, ResourceId,
)

async with YouTubeClient(access_token="TOKEN") as yt:
    # Create a playlist
    playlist = await yt.playlists.insert(
        body=Playlist(
            snippet=PlaylistSnippet(title="My SDK Playlist"),
            status=PlaylistStatus(privacy_status=PrivacyStatus.PUBLIC),
        ),
    )

    # Add a video to the playlist
    await yt.playlist_items.insert(
        body=PlaylistItem(
            snippet=PlaylistItemSnippet(
                playlist_id=playlist.id,
                resource_id=ResourceId(
                    kind="youtube#video",
                    video_id="dQw4w9WgXcQ",
                ),
            ),
        ),
    )

Comments

from youtube import (
    YouTubeClient, CommentThread, CommentThreadSnippet,
    CommentThreadPart, Comment, CommentSnippet, CommentPart,
)

async with YouTubeClient(access_token="TOKEN") as yt:
    # Post a top-level comment
    thread = await yt.comment_threads.insert(
        body=CommentThread(
            snippet=CommentThreadSnippet(
                video_id="VIDEO_ID",
                channel_id="CHANNEL_ID",
                top_level_comment=Comment(
                    snippet=CommentSnippet(text_original="Great video!"),
                ),
            ),
        ),
    )

    # List comments on a video
    async for thread in yt.comment_threads.iter_video_threads(
        "VIDEO_ID",
        parts=[CommentThreadPart.SNIPPET, CommentThreadPart.REPLIES],
    ):
        top = thread.snippet.top_level_comment.snippet
        print(f"{top.author_display_name}: {top.text_display}")

Analytics

from youtube import YouTubeClient

async with YouTubeClient(access_token="TOKEN") as yt:
    report = await yt.analytics.query(
        start_date="2025-01-01",
        end_date="2025-12-31",
        metrics=["views", "likes", "estimatedMinutesWatched"],
        dimensions=["day"],
    )
    for row in report.rows:
        print(row)  # [date, views, likes, minutes_watched]

Captions

from youtube import YouTubeClient, Caption, CaptionSnippet, CaptionPart, CaptionFormat

async with YouTubeClient(access_token="TOKEN") as yt:
    # List captions for a video
    captions = await yt.captions.list(
        parts=[CaptionPart.SNIPPET],
        video_id="VIDEO_ID",
    )

    # Download a caption track as SRT
    srt_data = await yt.captions.download(
        id=captions.items[0].id,
        tfmt=CaptionFormat.SRT,
    )

    # Upload a new caption track
    await yt.captions.insert_from_file(
        body=Caption(
            snippet=CaptionSnippet(
                video_id="VIDEO_ID",
                language="en",
                name="English",
            ),
        ),
        file_path="/path/to/captions.srt",
    )

API Namespaces

Namespace	Methods	Description
yt.videos	list, insert, update, delete, rate, get_rating, report_abuse	Video management & upload
yt.channels	list, update	Channel info & branding
yt.playlists	list, insert, update, delete	Playlist CRUD
yt.playlist_items	list, insert, update, delete	Manage videos in playlists
yt.comments	list, insert, update, delete, set_moderation_status	Comment replies & moderation
yt.comment_threads	list, insert	Top-level comments
yt.subscriptions	list, insert, delete	Subscription management
yt.thumbnails	set, set_from_file	Custom thumbnail upload
yt.search	list	Search across YouTube
yt.captions	list, insert, insert_from_file, update, download, delete	Caption track management
yt.channel_sections	list, insert, update, delete	Channel page layout
yt.channel_banners	insert, insert_from_file	Banner image upload
yt.watermarks	set, unset	Channel watermark
yt.analytics	query	YouTube Analytics reports

Authentication

The SDK does not handle OAuth flows. You must provide a pre-obtained token:

• access_token — OAuth 2.0 token for full read/write access
• api_key — Google API key for read-only public data

# OAuth (full access)
client = YouTubeClient(access_token="ya29....")

# API key (public read-only)
client = YouTubeClient(api_key="AIzaSy...")

# Both (OAuth takes priority, API key added to all requests)
client = YouTubeClient(access_token="ya29....", api_key="AIzaSy...")

Development

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Lint & format
ruff check .
ruff format .

# Type check
mypy youtube/

License

MIT License — see LICENSE for details.