Turn everyday conversations into AI fuel .

We collect and provide ethical, high-quality Japanese voice data through a user-participatory conversation platform, supporting corporate AI development.

For Business App Details

100k+hours

Collectible Voice Data Volume

APPICompliance

Legal Compliance & Protection

Native Voice

Unscripted Natural Conversations

Why is "natural voice data" needed now?

As AI evolves, the "quality" demanded from data is changing.

AI is leaving the screen and entering the real world

With advances in robotics and AI agents, the ability to "understand sound" is becoming increasingly important.

Until now, AI has evolved by learning from text and images on the web. However, "embodied AI" such as robotics and autonomous driving systems requires "hearing" — voice information — to communicate with humans. To supplement real-world information that visual and text data alone cannot capture, voice data containing realistic environmental sounds is becoming essential.

Training data is plateauing, and synthetic data misses "real-world noise"

Clean training data can be created, but it fails to capture real-world noise, environmental variation, and unpredictability.

Web text has already been exhausted for training. Moreover, AI trained on "overly clean data" like synthetic speech or studio recordings has its limits. It cannot function in the noisy, uncertain real world — this "gap between ideal and reality" is a growing challenge. What's needed now is "raw field data" that includes noise.

The essence of voice data is not "words" but "how they're spoken"

Tone, pauses, hesitation — information that can't be captured in text is the key to voice AI.

Current AI excels at processing "what was said" but struggles with understanding "how it was said." Accurately capturing voice-specific information — tone, intonation, speaking speed, emotional nuance — remains a challenge. To achieve human-like dialogue, voice data containing information beyond text is essential.

Clean voice data becomes a competitive advantage for AI

Amid tightening regulations, data with clear consent and provenance leads to trust and differentiation.

Countries worldwide are advancing legislation on AI data usage. In this climate, ensuring transparency in data provenance and rights is becoming more than just risk mitigation — it's a factor that influences corporate trust. Using "ethical data" with proper consent and clear collection processes is becoming a new competitive axis.

There's a reason our data is chosen

Everyday conversations, optimized for AI training.

Specialized in Natural Conversation

To improve LLM and speech recognition accuracy, "living conversations" without scripts are essential. We provide data that includes human-like nuances such as hesitations, laughter, and back-channel responses.

APPI Compliant

We obtain usage consent from all users. Personally identifiable information is securely removed, and only clean, commercially usable data is provided.

Diverse Demographics

Data can be categorized and filtered by age, gender, region, and other attributes.

AI-Ready Data Formats

Metadata including transcription, speaker diarization, and emotion tags is attached, ready for immediate use in AI development.

Custom Collection

We also support custom data collection, such as "conversations on specific topics."

Sustainable Ecosystem

Users' "fun" becomes corporate "value."
A sustainable and transparent data circulation process.

Talk

Enjoy calls on the app

Earn

Earn points

Process

Anonymize & tag

Deliver

Provide as AI training data

Connect by voice,
create value.

Unscripted "living conversations" are the most valuable data. Through the mechanism of "earn points just by talking," we've made it possible to collect real conversation data that was previously difficult through traditional recording methods.

Try the App

Ready to innovate?

We flexibly accommodate your data needs and budget.
Feel free to reach out.