Name "Processor ⚑" "Phonetic alphabet πŸ”€" "Insta-clone πŸ‘₯" "Emotion control 🎭" "Prompting πŸ“–" "Speech control 🎚" "Streaming support 🌊" "Voice conversion 🦜" "Longform synthesis πŸ“œ" Zonos CUDA IPA πŸ‘₯ "🎭 8-type" πŸ—£πŸ“– "🎚 speed, pitch, 🎭" xVASynth CPU / CUDA ARPAbet+ ❌ "🎭 4-type πŸ˜‘πŸ˜ƒπŸ˜­πŸ˜― per‑phoneme" ❌ "🎚 speed, pitch, energy, 🎭; per‑phoneme" ❌ 🦜 ~15s max GPT-SoVITS CUDA IPA πŸ‘₯ 🎭πŸ‘₯ ❌ "🎚 speed, stability" ❌ 🦜 MetaVoice-1B CUDA πŸ‘₯ 🎭πŸ‘₯ ❌ "🎚 stability, similarity" πŸ“œ IMS-Toucan CUDA IPA πŸ‘₯ ❌ ❌ "🎚 speed, stability; per-phoneme" XTTS CUDA ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "🎚 speed, stability" 🌊 ❌ LLaSA CUDA πŸ‘₯ 🎭πŸ‘₯ πŸ—£πŸ“– StyleTTS CPUs / CUDA IPA πŸ‘₯ 🎭πŸ‘₯ ❌ 🌊 πŸ“œ Kokoro CPUs / CUDA / ROCm IPA ❌ fusion ❌ ❌ "🎚 speed" 🌊 πŸ“œ HierSpeech++ ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "🎚 speed, stability" 🦜 E2/F5 TTS CUDA πŸ‘₯ 🎭πŸ‘₯ πŸ—£πŸ“– "🎚 speed" πŸ“œ / by ~30s MaskGCT CUDA πŸ‘₯ 🎭πŸ‘₯ ❌ "🎚 set duration" Fish Speech CUDA ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "🎚 stability" 🌊 πŸ“œ Parler CUDA ❌ πŸŽ­πŸ“– πŸ“– πŸŽšπŸ“– 🌊 OpenVoice CUDA ❌ πŸ‘₯ "🎭 6-type πŸ˜‘πŸ˜ƒπŸ˜­πŸ˜―πŸ€«πŸ˜Š" ❌ WhisperSpeech CUDA ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "🎚 speed" MeloTTS CPU / CUDA ❌ ❌ "🎚 speed" Matcha-TTS IPA ❌ ❌ ❌ "🎚 speed, stability" Pheme CUDA ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "🎚 stability" Piper CPU / CUDA / ROCm ❌ ❌ ❌ ❌ "🎚 speed, stability" ❌ ❌ TorToiSe TTS ❌ ❌ ❌ πŸ—£πŸ“– 🌊 Bark CUDA ❌ 🎭 tags ❌ πŸ“œ / by ~13s TTTS CPU / CUDA ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ Amphion CUDA ❌ fusion ❌ ❌ "🎚 speed" VITS/ MMS-TTS CUDA ❌ ❌ ❌ "🎚 speed" AI4Bharat EmotiVoice Glow-TTS MahaTTS Neural-HMM TTS OverFlow TTS pflowTTS RAD-MMM RAD-TTS Silero Tacotron VALL-E