NVIDIA Nemotron ASR Multilingual

Alibaba Fun-ASR Flash

Alibaba Cloud DashScope non-realtime speech recognition with multilingual transcription, punctuation, and sentence/word timestamps.

Voxtral Small 24B

Mistral AI’s original Voxtral Small 24B transcribes recorded speech in English, French, German, Spanish, Italian, Portuguese, Dutch, and Hindi, with automatic language detection. Add a context prompt for names or terminology. Accepts MP3, WAV, FLAC, OGG, and M4A recordings up to 10 minutes, with direct uploads up to 4 MB or audio URLs up to 25 MB. Returns a completed JSON transcript without timestamps, speaker diarization, or translation.

ElevenLabs Sound Effects V3

Create sound effects from text prompts up to 450 characters. Choose a duration from 0.5 to 30 seconds or let the prompt determine the length, adjust prompt influence, and generate seamless loops. Download MP3 or Opus audio.

Mureka Speech

Generate speech from up to 500 characters with five preset voices or an uploaded voice sample.

Mureka Region Edit

Rewrite a song region of at least three seconds with replacement lyrics, up to 3,000 characters.

Mureka Music Transcription

Transcribe music into a downloadable archive containing sheet music and timed lyrics. Audio must be 10 seconds to 10 minutes and at most 100 MB.