Product Overview
AI Multilingual Meeting Tool is a post-processing AI system that analyzes recorded meeting files (audio and video) fully offline and automatically generates multilingual minutes, subtitles, and speaker identification. It comprehensively transforms the post-processing of multilingual meetings in international business, delivering annual cost savings of JPY 1.95–3.15 million and a 95% reduction in time.
Conventionally, multilingual meetings involved challenges such as interpretation costs (JPY 1.2–2.4 million per year), meeting minutes creation time (1.5 hours per week), and security risks from external data transmission. This tool solves all these challenges through fully offline operation.
Key Features
Multilingual Speech Recognition
The high-accuracy speech recognition engine transcribes audio in 99 languages. Technical terms are accurately recognized via custom dictionaries.
Translation across 419 Languages
The large-scale multilingual translation engine supports translation between any of the 419 supported languages, covering over 99% of the world's population.
Speaker Identification
The speaker diarization engine automatically identifies who said what, contributing to structured meeting minutes.
Structured Meeting Minutes Generation
A large language model summarizes the content and extracts agenda items, decisions, and action items.
Multilingual Subtitle Generation
Adds multilingual subtitles to videos, enabling later review of meeting content in various languages.
Accuracy Improvement through Multi-AI Coordination
Complementary processing across multiple AI engines detects and corrects recognition errors.
Technical Specifications
| Speech Recognition | High-accuracy speech recognition engine – 99 languages supported |
|---|---|
| Translation Engine | Large-scale multilingual translation engine – 419 languages supported |
| Speaker Identification | Speaker diarization engine – high-accuracy speaker diarization |
| Meeting Minutes Generation | Large language model – structured document generation |
| Processing Method | Fully offline, post-processing (Zero external data transmission) |
| Multi-AI Integration System | Multi-AI coordination through sequential processing |
| License | All components are commercially usable (combination of OSS-compliant licenses) |
Pricing
- 99-language speech recognition
- 419-language translation
- Speaker identification
- Automatic meeting minutes generation
- Subtitle generation
- Custom dictionary
- Fully offline operation
- All features of Personal Edition
- Organizational customization
- Technical term dictionary customization
- Implementation support
- Annual email support
- All features of Enterprise Edition Lite
- Mac Studio M3 Ultra hardware included
- Initial setup service
- On-site implementation training
- Dedicated support (1 year)
- Custom development support
System Requirements
| Recommended Environment (Enterprise Edition Pro) | Mac Studio M3 Ultra (192GB unified memory, 8TB SSD) |
|---|---|
| Minimum Environment | Mac Studio M2 Max (64GB unified memory, 2TB SSD) or MacBook Pro M4 Max |
| Supported OS | macOS 14.0 (Sonoma) or later |
| Network | Not required (fully offline operation) |
| Input Formats | Audio: MP3, WAV, M4A / Video: MP4, MOV, MKV |
| Output Formats | Minutes: Markdown, Word, PDF / Subtitles: SRT, VTT |