xiaohu-video-translate: Auto-Subtitle Foreign Videos with AI

If you've ever wanted to watch a foreign-language video but found the subtitle options lacking, xiaohu-video-translate is a project worth bookmarking. This MIT-licensed Python tool from xiaohuailabs automates the entire pipeline — from downloading a video to burning Chinese subtitles directly into the file — with a single command.

With 535 stars and growing, it's clearly struck a chord with developers and content consumers alike.

What Problem Does It Solve?

Adding accurate Chinese subtitles to foreign-language video content is surprisingly tedious. You'd normally need to download the video, run it through a transcription service, translate the result, clean up the output, and then use a separate tool to embed the subtitles — often paying API fees at multiple steps along the way.

xiaohu-video-translate collapses all of that into one workflow. The project description sums it up well: "对 AI 说一句话，把外语视频自动配上中文字幕" — roughly, tell the AI in one sentence, and it automatically adds Chinese subtitles to your foreign-language video.

Key Features

Based on the repository description and project scope, the tool covers these core stages:

Download — Fetches video from a URL (likely using a tool like yt-dlp under the hood)
Transcribe — Converts spoken audio to text, handled locally with zero API cost
Translate — Translates the transcribed text into Chinese
Polish — Refines the translation for natural readability, not just literal output
Burn — Hardbakes the final subtitles directly into the video file

The emphasis on fully local transcription is a meaningful detail. Transcription APIs can get expensive fast, especially for long-form content. Running it locally means you can process as many videos as you want without worrying about per-minute billing.

Who Should Use This?

This tool is a good fit for several types of users:

Developers and researchers who regularly consume English, Japanese, or other foreign-language technical content and want quick Chinese subtitles without manual effort
Content creators and educators localizing video material for Chinese-speaking audiences
Hobbyists who want a self-hosted alternative to paid subtitle services
Teams building localization pipelines who want a reference implementation or starting point

It's worth noting that the tool is opinionated — it's specifically oriented toward generating Chinese subtitles. If your target language is different, you'd need to adapt the translation step.

Where AI APIs Come In

While transcription runs locally (a genuine cost advantage), the translation and polishing steps are where AI models do meaningful work. Depending on how the tool is configured, this is where a service like KodaAPI could slot in naturally. KodaAPI provides a unified API key that routes to OpenAI, Anthropic, Google Gemini, DeepSeek, and 100+ other models — so if you're experimenting with which model produces the most fluent Chinese translations or best handles domain-specific terminology, you can swap models without managing multiple credentials or billing accounts.

That said, many users will simply use the defaults and get solid results without touching the API layer at all.

Getting Started

The project is open source under the MIT License and written in Python, so setup should be straightforward for most developers.

Clone the repository: git clone https://github.com/xiaohuailabs/xiaohu-video-translate
Install dependencies (check the repo's README for the exact requirements)
Point the tool at a video URL and let the pipeline run

Head over to the GitHub repository for full setup instructions, configuration options, and usage examples.

Final Thoughts

xiaohu-video-translate is a practical, well-scoped tool that solves a real workflow pain point. It doesn't try to do everything — it does one thing end-to-end and does it locally where it matters most. For Chinese-speaking developers drowning in foreign-language video content, this is the kind of automation that quietly saves hours every week.

Repo: xiaohuailabs/xiaohu-video-translate · ★ 535 · MIT License