DolphinVoice is focused on the Japanese market, helping voice AI better understand Japanese and fit real enterprise workflows. Built around speech recognition and pronunciation assessment, it provides reliable API, SaaS, and smart device services for Japanese companies, educational institutions, and multilingual service environments.
Core focus
From developer APIs to SaaS and intelligent devices, DolphinVoice addresses practical needs across speech understanding, speech generation, pronunciation assessment, and multilingual communication.
APIs: From real-time recognition to pronunciation assessment
Real-Time Speech Recognition (Realtime-Transcription)
A WebSocket streaming API optimized for long, continuous voice interactions. It supports mixed Chinese, Japanese, and English speech, with speaker diarization, intelligent punctuation, filler word removal, custom vocabulary, forced replacement, and noise reduction.
Audio & Video Transcription
A high-speed offline service for batch transcription of audio and video files. It supports document and subtitle output, custom vocabulary, forced replacement, speaker identification, spoken-language smoothing, and Chinese-English or Japanese-English mixed speech. One hour of audio takes about 65 seconds on average.
Short Speech Recognition
A fast API for converting short audio into text, designed for conversational AI, voice commands, IVR recognition nodes, and automated outbound voicemail recognition that drives downstream workflows.
Text-to-Speech
Converts text into natural, expressive speech with multiple voice options, controllable emotions and speaking styles, and PCM, WAV, or MP3 output.
Speech Evaluation (SOE)
An AI pronunciation assessment service for Japanese and English speaking practice and online evaluation. It delivers multi-dimensional scores within milliseconds, identifies missing sounds and mispronunciations, and supports phoneme, sentence, and open-ended tasks.
SaaS and intelligent devices
NihongoScore
An AI-powered Japanese speaking assessment service for foreign employees. Based on CEFR, it evaluates six dimensions and generates a comprehensive report in 15–20 minutes, supporting screening, job placement, and long-term development tracking.
DolphinTeams
An intelligent meeting platform combining a 360° panoramic camera, AI speaker tracking, beamforming microphones, real-time multilingual translation, AI meeting minutes, and smart summaries.
TralingoTalk
An integrated AI translation product with more than 30 days of standby time, real-time translation for 140+ online and 10+ offline languages, two-way translation, real-time subtitles, an AI assistant, intelligent Q&A, and multiple AI models.
Company overview
| Item | Details |
|---|---|
| Brand | DolphinVoice |
| Operating Company | DolphinAI Inc. |
| Founded | December 2023 |
| Representative Director | Masahiro Asakura |
| Address | 32F, Shinjuku Nomura Building, 1-26-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo 163-0590, Japan |
| Contact Email | contact@dolphin-ai.jp / voice.contact@dolphin-ai.jp |
| Security Certifications | ISO 27001 / SOC 2 |
| Data Processing Region | Japan |
| Deployment Model | Cloud-based SaaS |
Use cases
- Meeting minutes, interview transcription, and call center quality monitoring
- Video and podcast subtitles, film dubbing, and online education content
- Cross-border business, travel, multilingual meetings, and Japanese assessment for global talent
