NEWS
mocomoco Launches High-Performance Speech Recognition AI 'mocoVoice API'!
mocomoco, Inc. is officially releasing the speech recognition cloud service 'mocoVoice API'. This service provides an API for individuals and businesses, utilizing next-generation speech recognition technology to deliver an unprecedented experience in terms of both accuracy and speed.
＜mocoVoice Features＞
mocoVoice enables transcription from audio. In addition to providing audio transcription, it also offers the following features:
Domain specialization via dictionary feature
Register custom dictionary words tailored to specific industries or specialized fields to improve accuracy.
Example: Registering industry terms like 'Myna health insurance card' enables highly accurate transcription.Phonetic-free dictionary registration
Features a dictionary function that does not require specifying phonetic readings. Example: When registering 'Myna health insurance card' as a dictionary word, there is no need to add the phonetic reading 'mai-na-ho-ken-sho'.Ultra-fast inference
A one-hour audio file can be transcribed in as little as 3 minutes.High-speed proofreading powered by ChatGPT
Automatically proofreads the text after speech recognition to deliver high-quality prose.
In addition to audio input data, speech recognition can also be performed directly from video data.
＜Target Users＞
The mocoVoice API can be used by everyone from individual users to enterprise clients. We offer multiple plans to meet a wide range of needs.
＜Pricing Plans (*1)＞
For the mocoVoice API pricing plans, you can choose between plans without data learning and plans with data learning.
Plans Without Data Learning
Your data will not be used to improve the service. Recommended for users who prioritize privacy protection.
Plan Name | Base Fee (JPY) | Transcription Fee (JPY/min) | Free Tier (min) |
|---|---|---|---|
Free (Up to 5 hours/month) | 0 | 0 | 300 |
Standard | 950 | 2.64 | 300 |
Pro | 4,980 | 2.23 | 300 |
Enterprise | 13,800 | 1.73 | 300 |
Plans With Data Learning
Your data will be anonymized and used to improve the service and enhance the accuracy of mocoVoice.
Base Fee (JPY) | Transcription Fee (JPY/min) | Free Tier (min) | |
|---|---|---|---|
Free (Up to 5 hours/month) | 0 | 0 | 300 |
Standard | 950 | 1.98 | 300 |
Pro | 4,980 | 1.23 | 300 |
Enterprise | 13,800 | 1.04 | 300 |
All plans include the proofreading feature.
＜Apply for Service Usage Here＞
To start using the mocoVoice API, please apply for usage via this Google Form. After creating an account, you can try out the speech recognition features immediately.
＜Adoption Track Record＞
Among companies that have already adopted the service, it is being utilized for meeting minutes tailored to the financial domain, contributing to more efficient meetings and faster information sharing. It also handles specialized terminology and abbreviations unique to the financial industry, achieving accurate transcription of meeting minutes.
＜Development Background＞
Currently, typical speech recognition services often struggle with specialized terminology used in fields like medicine and finance. Furthermore, competitor services often face challenges such as the inability to handle audio where Japanese and English are mixed. Additionally, the time-consuming nature and high costs of speech recognition are seen as problematic. We developed mocoVoice to solve these issues.
＜Technical Features＞
mocoVoice is based on OpenAI Whisper, which boasts the highest performance in the speech recognition industry, combined with mocomoco's proprietary dictionary algorithms and acceleration technologies. The mocoVoice dictionary algorithm makes it possible to improve recognition accuracy even without phonetic readings for specific words. Through mocoVoice's acceleration technology, we have achieved faster speech recognition than OpenAI Whisper (*2).
We utilize ChatGPT for text proofreading, devising ways to accurately reflect words where the reading and spoken audio differ.
＜Support & Operations Structure＞
We have prepared a documentation page for the mocoVoice API ( https://docs.mocomoco.ai/ ). In addition to sample programs that are easy to implement, we also provide technical support.
A dashboard page is available where you can manage usage fees, teams, and API keys.
＜Future Outlook＞
The mocoVoice API is scheduled to add multilingual support and speaker diarization features in the near future. We also plan to gradually roll out the mocoVoice Playground page, where you can easily try out mocoVoice.
Starting with the high-performance speech recognition AI 'mocoVoice', mocomoco, Inc. plans to deploy even more convenient AI services that closely align with user needs.
*1 The above fees are current as of December 4, 2024. Please check the official website for the latest pricing.
*2 The comparison target is the transformers version of openai/whisper-large-v3.
Related Pages
mocoVoice API Usage Application Form
https://forms.gle/YU2a4aeHgwtF8E8L7
mocoVoice API Documentation Page
https://docs.mocomoco.ai/
mocomoco, Inc. Website
https://www.mocomoco.ai/