VideoLingo

VideoLingo is a self-hosted video translation app for creators and educators who need bilingual subtitles or dubbed versions of their videos. It brings transcription, translation and subtitle timing into one browser interface, with dubbing as an optional output. The project is open source under Apache 2.0; a separate hosted service offers subtitle translation and dubbing.

The app runs on Windows, macOS and Linux, with Docker support for Linux NVIDIA deployments. Its default speech recognition uses Qwen3-ASR and Qwen3-ForcedAligner locally. Apple Silicon uses MLX, while Intel Macs use CPU recognition. Translation connects to an OpenAI-compatible API. Optional cloud recognition through ElevenLabs or MAI-Transcribe-2 sends audio to the selected provider, so running the app yourself doesn't make every stage local.

Subtitle work includes word-level alignment, sentence segmentation and length limits for readable lines. Custom and AI-generated terminology help keep specialist vocabulary consistent, while optional translation review and rewriting address phrasing. It exports subtitle files and can produce videos with subtitles or dubbed audio. YouTube downloads use yt-dlp.

Dubbing supports GPT-SoVITS, OpenAI and Edge TTS; available languages depend on the chosen backend. Quality and synchronization depend on the audio, translation and voice service. The workflow doesn't automatically give each speaker a separate voice.

A local HTTP API lets scripts and AI agents process videos, check progress and retrieve outputs. Tasks can pause or resume, and completed work can survive retries.