VideoLingo

VideoLingo is a video translation app for creators and educators producing bilingual subtitles or dubbed videos. Its self-hosted browser interface brings transcription, translation and subtitle preparation into one workflow, though the selected providers determine which processing stays local.

• Timed subtitles: Transcribe speech with word-level alignment, split sentences into readable lines and export subtitle files or subtitled videos.
• Translation terminology: Use custom or AI-generated terminology for specialist vocabulary, with optional translation review and rewriting.
• Optional dubbing: Produce dubbed videos through GPT-SoVITS, OpenAI or Edge TTS; language availability depends on the voice backend.
• Automation: Process videos through a local HTTP API for scripts and AI agents, with task controls and progress resumption.

The Apache 2.0 app runs on Windows, macOS and Linux. Docker deployment supports Linux with NVIDIA GPUs. Default recognition uses Qwen3-ASR and Qwen3-ForcedAligner locally, with MLX on Apple Silicon and CPU recognition on Intel Macs.

Translation requires an OpenAI-compatible Chat Completions API that returns the workflow's structured output. Optional ElevenLabs or MAI-Transcribe-2 recognition sends audio to that provider. A separate hosted service provides subtitle translation and dubbing.

Mixed-language speech can produce inaccurate text or timing. Dubbing doesn't automatically assign each speaker a separate voice, and voice quality and synchronization depend on the translation and TTS service.