Research

A five-model multilingual speaker-identification benchmark led by identity-disjoint held-out evaluation, with clean and 900-speaker supporting checks.

Introducing VoicePing Diarization v0.1, a multilingual speaker diarization model evaluated on a 42-file benchmark with NeMo, pyannoteAI precision-2, AssemblyAI, and Deepgram.

Introducing VoicePing MT v0.1 in a 100-row English-to-Japanese benchmark with GPT-5.5 quality judging, observed latency, and comparisons against DeepL, Sakana Translate, GPT-5 mini, Google Translate, Qwen, Azure, and Llama.

Introducing VoicePing ASR Model V0.1, an Asian-language-focused speech-to-text model for Japanese, Korean, Chinese, Vietnamese, and English.

Benchmarking five emotional text-to-speech models for Japanese and Chinese across six target emotions, with SenseVoice emotion recognition, emotion2vec anchors, CER, naturalness, runtime, and listening examples.

Open-source cross-platform mobile app for fully offline speech translation — combining on-device ASR (SenseVoice), neural machine translation, and TTS on iOS and Android with system audio capture