Skip to content

User Guide ​

Quick Start ​

  1. On the main screen, click "New Task".
  2. Paste a video URL (YouTube / Bilibili, etc.).
  3. Choose the target action: Download only / Download + Transcribe / Download + Transcribe + Rewrite / Full pipeline (including publishing).
  4. Click "Start". Creator runs each stage in sequence; track real-time progress in the task panel.

Content Pipeline in Detail ​

Stage 1

Video Download

Supports 1000+ video platforms (see the full list of sites supported by yt-dlp).

See the full list at sites supported by yt-dlp.

Common sources:

PlatformURL Format
YouTubehttps://www.youtube.com/watch?v=xxx
Bilibilihttps://www.bilibili.com/video/BVxxx
Douyinhttps://www.douyin.com/video/xxx
Twitter/Xhttps://x.com/user/status/xxx

Download options:

OptionDescriptionDefault
Video qualityBest / 1080p / 720p / Audio onlyBest
SubtitlesAutomatically download available subtitlesOn
ProxyHTTP/SOCKS5 proxy (for accessing overseas platforms)None
Stage 2

Audio Transcription

Converts audio to text using the OpenAI Whisper model.

ModelSizeSpeedAccuracyBest for
tiny75 MBVery fastFairQuick preview
base142 MBFastGoodEveryday use (default)
small466 MBMediumExcellentWhen higher accuracy is needed
medium1.5 GBSlowOutstandingProfessional content, multilingual

Language support: Detects the language automatically, or specify it manually (supports 99 languages including Chinese, English, and Japanese).

Stage 3

AI Rewriting

Processes the transcribed text by calling AI models through the Lurus API.

Process the transcribed text by calling AI models through the Lurus API.

Rewrite modes:

ModeDescriptionBest for
TranslateTranslate foreign-language content into ChineseLocalizing overseas videos
PolishFix grammar and refine wordingImproving copy quality
CondenseExtract core points and shorten the lengthLong video → short copy
ExpandAdd detail and backgroundShort content → long article
Style transferAdjust tone (formal ↔ casual)Adapting to different platforms
SEO optimizationGenerate titles, tags, and summariesImproving search rankings

Model selection:

Recommended modelBest forCost
deepseek-chatChinese rewriting, translationLow
gpt-4oEnglish content, complex rewritesHigh
claude-3-5-sonnetCreative writing, long textMedium
Stage 4

Auto-Publish

Publishes automatically via a chromedp headless browser.

Only 3 platforms support auto-publishing

Currently only WeChat Official Account / Douyin / Xiaohongshu support auto-publishing. For other platforms (YouTube Shorts / TikTok / Instagram Reels, etc.), copy can be generated but must be published manually.

First-time setup:

  1. Settings → "Publishing Platforms" → select the target platform.
  2. Click "Log in" to open a browser window.
  3. Log in manually via QR code / password; Creator saves the login state.

Publishing settings (configured independently per platform):

SettingDescription
Title templateSupports variables: , ,
TagsAuto-generated or manually set
CoverAuto-captured from the video or manually uploaded
Publish timePublish immediately or schedule

Batch Processing ​

  1. Create a text file with one video URL per line.
  2. In Creator, click "Batch Import" and select the file.
  3. Choose a unified processing configuration.
  4. Click "Start All".

Task Status ​

Each task is tracked independently:

StatusDescription
queuedWaiting to be processed
downloadingDownloading the video
transcribingTranscribing speech
rewritingAI is rewriting
publishingAuto-publishing
completedFully complete
failedA stage failed (can retry)

Resume from checkpoint

A failed task can restart from the failed stage—no need to reprocess from the beginning.

Budget Control ​

Avoid consuming too many tokens during the AI rewriting stage:

SettingDescriptionDefault
Per-task token limitMaximum tokens a single task may use10,000
Daily total limitMaximum tokens consumed per day100,000
Over-limit behaviorPause / Skip rewriting / NotifyPause

Configure under Settings → "Token Budget".


Keyboard Shortcuts ​

ShortcutFunction
Ctrl+N / Cmd+NNew task
Ctrl+V / Cmd+VPaste URL and create a task
Ctrl+Shift+S / Cmd+Shift+SOpen settings
SpacePause/resume the current task
DeleteDelete the selected task

Troubleshooting ​

Video download fails?

Check your network (overseas videos may need a proxy); some platforms have anti-scraping measures—update Creator to the latest version; confirm the URL format.

Transcription is inaccurate?

Use a larger Whisper model (Settings → Whisper Model); manually specify the audio language; heavy background noise reduces accuracy.

AI rewriting times out?

Check your API Key balance; split overly long text into segments; switch to a faster model (such as deepseek-chat).

Publishing fails?

The platform login may have expired—scan the QR code again; check whether the content violates platform rules (title/images/sensitive words); platform API changes may require waiting for a Creator update.


Next Steps ​

Next steps

Powered by Lurus Technologies