KinaBot is a dignity-first, privacy-aware AI system for longitudinal speech reflection, healthy aging, and family-centered care.
KinaBot helps people reflect on observable speech and communication patterns across repeated, natural-language samples. It is designed to support personal awareness, family connection, and more informed conversations with care professionals while preserving the older adult's autonomy.
KinaBot is not a medical device, not a diagnostic tool, and not a replacement for licensed healthcare professionals.
KinaBot is currently maintained by Aoi Minamoto through AImoji LLC.
Earlier exploratory materials may remain in this repository for transparency and project history. They do not indicate current ownership, employment, maintainership, or product rights.
Current Aoi-maintained development is located in aoi_kinabot_app/.
This repository preserves earlier exploratory work and product-direction discussions for transparency. Current development focuses on a new KinaBot implementation maintained by Aoi Minamoto through AImoji LLC. Future public-facing releases will be developed separately from earlier exploratory materials, with clearer safety, privacy, and responsible-use boundaries.
Current Aoi-maintained development is in aoi_kinabot_app/.
KinaBot welcomes code, tests, documentation, translation, accessibility review, privacy review, reproducible research, issue triage, and product feedback.
- Contributing guide
- Code of Conduct
- Security and private vulnerability reporting
- Product and engineering roadmap
- Release process
- Current application changelog
- Licensing guide and commercial licensing
- Community/commercial boundary
- Contributor agreement, trademark policy, and authorship provenance
Start with a scoped issue
or review tasks labeled good first issue.
KinaBot V1 does not use cognitive age, biological age, dementia risk, medical risk labels, or diagnosis-like composite scores. V1 focuses on separate speech and language feature scores, trend reflection, consent-first access, and privacy-aware data handling.
Please treat files outside aoi_kinabot_app/ as historical exploratory materials unless otherwise stated.
The historical materials remain public to preserve provenance and show the project's continuing technical and responsible-design evolution. The current changelog, decision log, and Git history document when major design boundaries changed and why.
KinaBot is built around a simple belief: technology for aging should protect human dignity, strengthen family connection, and help people seek support earlier without fear or stigma.
The project focuses on:
- Dignity first: preserve each person's agency and avoid reducing anyone to a score or label.
- Privacy aware: minimize sensitive voice and identity data throughout the product lifecycle.
- Longitudinal reflection: describe patterns across time instead of drawing conclusions from a single sample.
- Healthy aging: support everyday reflection and constructive conversations, without diagnosis or disease-risk claims.
- Family-centered care: help people share understandable observations with trusted family members, caregivers, and professionals on their own terms.
The official product name is KinaBot. Kina is the short name used throughout the project. The project was originally launched under the name Kizuna in May 2025, and some early public references may still use that name.
The current Aoi-maintained V1 direction focuses on:
- Record short speech samples
- Convert speech to text
- Analyze basic language patterns such as word variety, sentence structure, speaking pace, and emotional tone
- Generate a simple report for personal reflection or family discussion
- Support English and Japanese speech recognition in the current interface
KinaBot is focused on healthy aging and family-centered care. Manufacturing, workplace monitoring, employee assessment, and employment decision-making are outside the product's scope.
KinaBot does not:
- Diagnose dementia, cognitive impairment, or any medical condition
- Replace doctors, clinicians, therapists, or professional cognitive assessment
- Provide emergency support
- Guarantee that a speech pattern means a health condition exists
- Determine medical risk without clinical review
If you or a family member has health concerns, please consult a qualified healthcare professional.
KinaBot works with sensitive voice and language data. Users should only record or upload speech when they have proper consent from every person whose voice may be included.
Current V1 design direction:
- Audio may be processed for speech-to-text transcription.
- The current local Python implementation uses the
SpeechRecognitionlibrary and may use Google speech recognition services unless replaced by a local or private transcription backend. - In local development, recordings may be temporarily saved during analysis.
- Generated reports may include transcripts and speech pattern summaries.
Before using KinaBot with real users, review Privacy.md, Medical-Disclaimers.md, and LEGAL-DISCLAIMER.md.
KinaBot should be used with:
- Clear user consent
- Respect for older adults' autonomy
- Minimal data collection
- Transparent explanation of what is analyzed
- Human review before any care decision
- Professional medical consultation when concerns arise
KinaBot is an early-stage source-available public-benefit project for research, education, and non-commercial exploration. The project is not clinically validated and should not be used as a standalone medical screening system.
This repository currently uses a custom non-commercial source-available license. Personal, academic, and non-profit research use is allowed under its terms. Commercial use requires prior written permission. Because the license restricts fields of use, it is not an OSI-approved open-source license; please describe the current project as source-available, not OSI open source.
See LICENSE.md for details.