Transform audio and video into accurate text, translated subtitles, and natural dubbed voiceovers. WaveCap is an all-in-one local AI studio built for macOS—giving you complete control over transcription, translation, and speech generation without sending data to the cloud. KEY HIGHLIGHTS * Up to 4x Faster Inference: Native Apple Silicon and MLX runtime optimizations deliver extreme local performance. * 20+ Transcription Models: Choose from Whisper, Parakeet, Qwen3 ASR, Cohere Transcribe, Voxtral, and more. * Translate to 50+ Languages: Translate subtitles and transcripts using local LLMs like Qwen, Gemma, and Llama. * Timeline-Aligned AI Dubbing: Generate realistic voice tracks matched to your video's timing for fully localized output. * 50+ State-of-the-Art Models: Test top-tier open models for noise isolation, ASR, translation, diarization, and TTS. 100% LOCAL & PRIVATE All processing runs directly on your Mac. Download your models once, then run noise reduction, transcription, translation, and dubbing with zero cloud reliance, no subscription fees, and no active internet connection. COMPREHENSIVE PIPELINE 1. Smart Import: Recursive folder scanning, mono conversion, and automatic language detection. 2. Voice Isolation: Clean noisy audio using DeepFilterNet and Demucs. 3. Precision STT: Transcribe complex multi-speaker recordings accurately. 4. AI Translation & Polishing: Fix grammar and translate text locally. 5. Speaker Diarization: Label up to 4 distinct speakers with Sortformer. 6. Localized Dubbing: Create time-aligned voice tracks with optional background audio recovery. BATCH WORKFLOW & EXPORTS * Queue Processing: Handle single files or entire folder trees in parallel. * Subtitle & Text: Export SRT/VTT (with speaker tags), TXT, CSV, or JSON. * Audio & Video: Export standalone voice tracks (WAV/M4A) or fully dubbed MOV videos. SUPPORTED AI MODEL FAMILIES * Speech-to-Text (STT): Whisper, Parakeet, Qwen3 ASR, Cohere Transcribe, MOSS Transcribe, SenseVoice, Voxtral, GLM-ASR, Nemotron ASR * Vocal Processing & Cleanup: DeepFilterNet, Demucs / MDX * LLM Text Polishing & Translation: Gemma, Qwen, Nemotron, Llama * Speaker Diarization: Sortformer * Text-to-Speech (TTS): Qwen3 TTS, OmniVoice, Breeze TTS * Voice Detection & Language ID: Silero VAD, ECAPA-TDNN VoxLingua107
Photo & Video · iPhone, iPad · Free
