VideoLingo is an open-source, one-click video localization tool that turns foreign-language videos into subtitled or fully dubbed versions, aiming for Netflix-level subtitle quality. It chains together speech recognition, subtitle segmentation, AI translation, terminology handling and text-to-speech dubbing in a single Streamlit app. Created in 2024, it has earned about 18,900 stars on GitHub and is still actively developed in 2026.
The pipeline starts by downloading a YouTube video with yt-dlp, then transcribes it with word-level alignment using Qwen3-ASR and a forced aligner. NLP and LLM steps cut the transcript into natural subtitle lines, build a custom terminology glossary for consistent names and terms, and translate with optional reflection and rewriting passes for natural phrasing. Finally, VideoLingo can burn in dual subtitles or generate a dubbed track with voice options such as GPT-SoVITS voice cloning, CosyVoice, OpenAI or Edge TTS. Input is supported for English, Chinese, Japanese, French, German, Spanish, Italian and Russian, and an HTTP API plus pause and resume controls make batch jobs manageable.
VideoLingo is completely free and open source under the Apache-2.0 license. You run it locally and only pay for any cloud LLM, ASR or TTS APIs you choose to plug in.
VideoLingo is for creators, translators and fansub-style teams who repost or localize videos across languages and want professional subtitles and dubbing without manual, line-by-line work.