AIdeas: MUSUBiAi - IoT Family Praise Platform
Busy parents want to praise their children but lack specific information and skills to do so effectively. MUSUBiAi solves this by capturing children's achievements in real-time and providing AI-powered praise recommendations, strengthening family bonds and nurturing self-esteem.
"Busy parents miss their children's achievements. MUSUBiAi captures those moments and AI teaches you how to praise specifically, nurturing children's self-esteem and deepening family bonds."
App Category
Daily Life Enhancement
MUSUBiAi is an IoT platform that aims to improve children's self-esteem by activating communication between family members and forming a bidirectional feedback loop centered on children's achievements.
Team Name: ONIGIRi
Team member: @yujishirai, @tanuki
My Vision
The Importance of Nurturing Children's Self-Esteem
In children's healthy development, self-esteem is extremely important. Self-esteem is the sense of "I am a valuable person" and "I am loved," and one of the most effective ways to nurture this is for parents to praise their children specifically.
Characteristics of children with high self-esteem:
- High motivation to try new things
- Can view failures positively
- Can maintain good relationships with others
- Have the strength to overcome difficult situations
Risks when self-esteem is not nurtured:
- Tend to measure self-worth by SNS "likes"
- May engage in problem behaviors to attract attention
- Tend to fear failure and stop trying new things
- May lack confidence in themselves and narrow their future possibilities
However, modern families face barriers that prevent them from sufficiently practicing this act of "praising".
Target Users
Children: Ages 5-12 (elementary school)
- Age when they can recognize their achievements and express them in words
- Generation familiar with digital devices
- Critical period for developing self-esteem
Parents: Wide age range (late 20s to 50s)
- Dual-income busy families
- Want to value communication with children but have time constraints
Barriers to Nurturing Self-Esteem Despite the Desire
In modern nuclear families, both parents are busy working, and they increasingly miss their children's small achievements. Children want to tell their parents about achievements like "I got 100 points on my math test today," "I made up with my friend," or "I learned to do a pull-up," but by the time their parents come home, the excitement and detailed context of that moment have faded.
Parents also have a strong desire to "praise their children more" and "build their self-esteem," but when they come home tired from work and their child says "Today, something happened," they don't know the specific situation and can only respond with generic words like "That's great" or "Good job." Furthermore, many parents struggle with concerns like "I don't know how to praise" or "I always use the same words."
What MUSUBiAi Aims to Achieve
MUSUBiAi aims to use technology to support parents in giving specific praise to their children and strengthen family bonds. The system is not just a recording tool, but functions as a "catalyst" to activate communication within the home and nurture children's self-esteem.
MUSUBiAi's Approach
MUSUBiAi activates communication within the home by forming a bidirectional feedback loop centered on children's achievements.
Basic Flow:
- Child: Record voice at the moment of achievement (preserving emotions and details vividly)
- System: Convert voice to text, AI analyzes and recommends specific praise to parents
- Parent: Praise children in their own words, referring to recommendations
- Feedback Loop: By repeating this cycle, children's self-esteem is nurtured and motivation for the next achievement increases
Key Points:
- System provides "hints" not "answers," allowing parents to actively engage with their children
- Recording at the moment of achievement solves information loss and lead time problems
- AI generates personalized recommendations considering child's profile (age, interests, past achievements)
- Parents can learn and grow "praising skills"
Origin of the Name MUSUBiAi
The name MUSUBiAi comes from the Japanese word "Musubi (結び)."
"Musubi" is a traditional Japanese concept with the following meanings:
- Connection: Bonds between people
- Mediation: Acting as a bridge to connect and bring together things that are apart
- Fruition: Efforts bearing fruit
This project uses technology to "musubi" family bonds, nurtures children's self-esteem through parents' specific praise, and supports children's efforts to "bear fruit." By combining "AI," it fuses modern technology with traditional Japanese values.
Why This Matters
Why isn't children's self-esteem sufficiently nurtured in modern homes despite understanding its importance? Why can't parents give specific and meaningful praise even though they want to praise their children?
Hypothesis: Home Communication Feedback Loop is Not Functioning
We hypothesize that in modern homes, the communication feedback loop is partially dysfunctional. This is not a problem of parental love or motivation, but structural barriers exist.
Structural Issues Children Face: Information Loss and Lead Time
Issue 1: Information Loss
Children have the desire to "be praised," but when conveying to parents, the emotions and specific context of that moment are lost.
Specific example:
- Moment of achievement: "Yes! I got 100 on my math test! I solved the difficult problems! I'm happy!"
- When telling parents: "Today, I got 100 on my test" (emotions have faded)
- Parent's response: "That's great" (don't know the specific situation, generic praise)
Issue 2: Lead Time Problem
Due to time delay from achievement to telling parents, the excitement and details of that moment are lost.
Timeline example:
- 14:00 - Receives test results at school (excitement MAX)
- 15:30 - Returns home (still happy but other things on mind)
- 18:00 - Parents come home ("Oh right, there was a test today" level of memory)
- 19:00 - Remembers and tells at dinner (emotions have considerably faded)
This lead time causes the "freshness" of the child's achievement to be lost, and parents cannot grasp the specific situation.
Structural Issues Parents Face: Lack of Information and Praise Skills
Issue 1: Lack of Information
Parents want to motivate their children to take on more challenges, but they are in a situation where they cannot fully grasp detailed information and feelings.
Parents' inner voice:
- "Getting 100 on a test is great, but I don't know which problem was difficult"
- "I want to know specifically how hard they studied"
- "How does it compare to other children, how does the child feel about it"
Due to lack of information, parents cannot give specific and meaningful praise and end up with generic words like "That's great" or "Good job."
Issue 2: Insufficient of Praise Skills
Many parents understand the importance of praising children, but have not acquired the skill of appropriately motivating praise on the spot.
Common parent concerns:
- "I don't know how to praise"
- "I always use the same words"
- "Will too much praise spoil them?"
- "What specifically should I praise?"
Due to this lack of praise skills, parents cannot sufficiently recognize their children's achievements and provide appropriate feedback.
Dysfunction of Existing Feedback Loop
Due to the above challenges, the home communication feedback loop falls into dysfunction as follows:
- Child → Achievement (excitement MAX)
- Lead Time → Information loss, emotional decay
- Parent → Receives incomplete information
- Parent → Insufficient positive feedback skills, includes criticism etc while praising
- Child → Decreased self-esteem, weakened emotional connection
- Result → Loss of motivation for next achievement
Specific example:
- Child: "I won a science award!"
- Parent: "Well done... but your room still isn't cleaned up"
- Child: (I was praised, but also criticized...)
By repeating this loop, children's self-esteem is eroded, the emotional connection between parent and child weakens, and the willingness to try new things is lost.
How to Nurture Self-Esteem: Points Shown by Research
According to developmental psychology research, what's important in nurturing children's self-esteem is not "how much to praise" but how and who is involved.
1. Praise Specifically (Descriptive Praise)
Praise that specifically verbalizes behavior has been shown to enhance children's self-evaluation more than abstract praise.
- ❌ "That's great!"
- ✅ "You solved it without giving up until the end"
- ✅ "You drew it by devising it yourself"
By making clear what was good, children develop the ability to find value in their own actions.
2. Don't Exaggerate, Recognize the Process
Focusing on process such as effort and ingenuity nurtures the growth mindset that "you can grow if you work hard." On the other hand, excessively elevating children as "more special than other children" has been reported to potentially increase narcissism. Reality-based involvement rather than exaggeration is important.
3. Feedback from Parents
Research shows that warm and stable involvement from parents (parental warmth) nurtures children's self-esteem. Self-esteem is nurtured not from the correct words themselves, but from words exchanged within a trusting relationship.
4. Continuous Involvement
Not one-time praise, but continuous interest and dialogue supports stable self-esteem. Praise focused on process has also been shown to promote children's exploratory behavior and willingness to challenge.
Reference Research:
- E Brummelman, C Sedikides (2020). Raising Children With High Self-Esteem (But Not Narcissism). Child Development Perspectives, 14(2), 83–89.
- JM Robichaud, F Grenier, M Joussemet, GA Mageau (2024). The Role of Descriptive and Non-Specific Outcome-Oriented Praise in Child Self-Esteem: A Multiphase, Multimethod Investigation. Journal of Child and Family Studies.
- E Brummelman, S Grapsas, K van der Kooij(2022). Parental Praise and Children's Exploration: A Virtual Reality Experiment. Scientific Reports.
MUSUBiAi's Solution
MUSUBiAi uses technology to solve the above challenges as follows:
1. Prevent Information Loss with Real-time Recording
- Recording by voice at the moment of achievement preserves emotions and details vividly
- Parents can hear the child's "raw voice" and grasp the situation specifically
2. Support Praise Skills with AI Assistance
- AI analyzes achievement content and recommends specific and appropriate praise
- Parents learn "how to praise" and improve praise skills
- Recommendations are based on research: "specific," "process-focused," "not exaggerated"
3. Visualize Continuous Growth with Summaries and Profiles (Phase 3 in Development)
- Grasp children's achievements chronologically with daily, weekly, and monthly summaries
- Chronicler Agent (summary generation AI) compresses achievement records and creates growth narratives
- Visualize achievement trends like "worked hard on math 3 times this week" or "many friendship-related achievements last month"
- Parents can look back on their child's growth trajectory and recognize continuous effort processes
- Becomes valuable data as a growth record that can be reviewed in the future
- Profiler Agent (profile analysis AI) automatically generates child profiles (interests, strengths, growth patterns) from achievement history
- Analyzes child characteristics like "good at math" or "proactive in building relationships with friends"
- By combining profiles and summaries, proposes more personalized, reality-based praise
- Enables praise based on specific facts like "You've been working hard on math last week too. Your attitude of steady effort is wonderful"
- Parents can deeply understand their children and give specific encouragement tailored to their interests and strengths
4. Reconstruct Feedback Loop
- System functions as a "catalyst" to activate home communication cycles
- Children learn "when I record achievements, my parents praise me" and motivation for next achievement increases
MUSUBiAi is designed to satisfy all four points shown by research (specifically, without exaggeration, parent-led, continuously). What we aim for is not "praise that momentarily lifts mood" but formation of self-esteem that accumulates within relationships.
How I Built This
System Operation Flow and Roles
MUSUBiAi realizes a bidirectional feedback loop where children, system, and parents collaborate.
1. Child-Initiated: Recording the Moment of Achievement
Children aged 5-12 press a button on the device and record their voice at the moment they achieve something.
Examples of recorded content:
- "Today, I got 100 points on my math test! I solved the difficult problems too!"
- "I had a fight with my friend Kenta, but I apologized first and we made up"
- "I did a pull-up for the first time! My teacher praised me too!"
Key Points:
- Recording at the moment of achievement preserves emotions and details vividly
- Speaking in their own words helps children reflect on achievements and deepen self-awareness
- Simple two-button operation allows children to use it independently
2. System's Responsibility: Organizing Information and Supporting Parents
The system converts the child's voice to text, and AI analyzes the content to provide appropriate praise recommendations to parents.
Information provided by the system:
- Summary of achievement (what was accomplished)
- Background of achievement (what effort was involved)
- AI-recommended praise (specific words that promote child's growth)
- Connection to past achievements ("You've been working hard on math last week too")
Key Points:
- The system provides "hints" for parents to praise in their own words, not "answers"
- AI generates personalized recommendations considering the child's profile (age, interests, past achievements)
- Supports parents in giving specific and meaningful praise even when busy
3. Parent's Responsibility: Praising Children in Their Own Words
Parents refer to the system's recommendations while praising and complimenting their children in their own words.
Examples of parent actions:
- Receive Slack notification at work and understand the achievement (mental preparation)
- Wait for the child to speak up after coming home
- When the child says "Today, I got 100 points on my math test," ask specifically "That's amazing! Which problem was difficult?"
- Prioritize drawing out the child's words while utilizing information from the system
- Encourage at dinner time: "It's the result of studying hard every day. I'm looking forward to your next test"
Key Points:
- The system is only "support," and it's important for parents to actively engage with their children
- By referring to recommendations, parents can learn "praising skills" and grow themselves
- Children feel "my parents are watching me" and gain a sense of security
4. Feedback Loop: Continuous Communication Cycle
By repeating this series of flows, a continuous communication cycle is formed within the home.
Effects of the feedback loop:
- Children learn "when I record achievements, my parents praise me" and motivation for the next achievement increases
- Parents can visualize their child's growth and give specific praise
- Communication throughout the family becomes more active and bonds are strengthened
- Children's self-esteem is nurtured and their willingness to take on future challenges increases
Data Flow Diagram
This diagram shows the flow of information from the user's perspective (from child's achievement recording to parent notification)
MUSUBiAi Architecture Overview
This diagram shows technical implementation details (AWS service configuration and integration)
Key AWS Services and Selection Reasons
Device Layer
- AWS IoT Greengrass V2: Execute voice processing with edge computing
- Privacy-focused: Text conversion on device without sending voice data to cloud
- Offline support: Operates during network disconnection and auto-syncs on reconnection
- Amazon ECR: Container image management for Greengrass components
- Store container images for device applications
- Efficient version management and deployment
Backend Layer
- AWS IoT Core: Secure communication between devices and cloud
- Authentication with X.509 certificates
- Low-latency communication with MQTT protocol
- Amazon SNS + Amazon SQS: Fan-out architecture
- Distribute messages from IoT Core to multiple subscribers
- Design that easily accommodates future feature expansion
- AWS Lambda: Serverless processing
- Scalable with low operational costs
- Amazon DynamoDB: Data storage
- Store achievement records and family data
- Flexible scaling with on-demand capacity
- Amazon S3: Context storage
- Child profile information
- Achievement summary context
- Reference data for AI praise generation
- Amazon Bedrock Agent (The Bridge): AI praise generation and orchestration
- Foundation Model: Uses Claude Sonnet 4.5 (latest high-performance AI model)
- Generates personalized praise recommendations based on child's achievement content
- Integrates with Knowledge Base to reference family context (profiles, past achievements)
- Executes Slack notifications through Action Groups
- Supports parents' praise skills
- Real-time processing (activates immediately upon achievement recording)
- Amazon Bedrock (Chronicler/Profiler Agents - Phase 3): Summary generation and profile analysis
- Chronicler Agent: Automatically generates daily, weekly, and monthly summaries, visualizing children's growth records chronologically
- Grasp achievement trends like "worked hard on math 3 times this week" or "many friendship-related achievements last month"
- Parents can look back on their child's growth trajectory and recognize continuous effort
- Profiler Agent: Automatically generates child profiles (interests, strengths, growth patterns) from achievement history
- Analyzes child characteristics like "good at math" or "proactive in building relationships with friends"
- By combining profiles and summaries, proposes more personalized, reality-based praise
- Enables praise based on specific facts like "You've been working hard on math last week too. Your attitude of steady effort is wonderful"
- Direct Bedrock API calls from Lambda functions (simple batch processing)
- Amazon EventBridge: Periodic execution trigger
- Real-time: Activates Bridge Agent upon achievement recording
- Periodic execution: Activates Chronicler Agent for weekly summary generation (to be implemented in Phase 3)
- Schedule-based automation processing
UI Layer
- Slack Integration: Rapid MVP development
- Implemented in Phase 1
- Delivers information to parents instantly with real-time notifications
- Web App (Phase 4): Planned
- Amazon Cognito: User authentication and authorization
- Secure user management
- Social login support
- Amazon CloudFront: Content delivery
- High-speed delivery with global CDN
- HTTPS communication optimization
- Amazon API Gateway: RESTful API
- Connection between backend and frontend
- API throttling and security
- Dashboard, analytics features, achievement timeline
- Provide richer user experience
Development Phases and Progress
Phase | Focus | Key Features | Status |
|---|---|---|---|
Phase 1 | MVP | Device (Button → Voice → Text) → AWS → Slack notification | ✅ Complete |
Phase 2 | AI Features | Bridge Agent integration, AI praise generation (real-time) | ✅ Complete |
Phase 3 | Engagement | Chronicler Agent (summary generation), Profiler Agent (profile analysis), User feedback collection | 🚧 In Progress |
Phase 4 | Refinement | Web app, dashboard, user feedback response | 📋 Planned |
Current Progress: Phase 2 complete, Phase 3 partially implemented (Chronicler/Profiler Agents in development)
Key Technical Decisions
- Voice Processing at Edge: Privacy-first design that doesn't send voice data to cloud
- Offline Queuing: Continues operation during network disconnection, ensuring reliability (max 100 items, 24-hour buffer. Deletes oldest data when exceeded)
- Slack Integration: Built MVP quickly in Phase 1-2 to gain early user feedback
- Fan-out Architecture: Design that easily accommodates future feature expansion (email notifications, mobile push, etc.)
- Bedrock Agent Adoption: Implements complex AI workflows concisely with managed orchestration, Action Groups, and Knowledge Base integration
Security and Privacy
MUSUBiAi implements multi-layered security measures to protect families' important data.
Device Authentication
- Mutual TLS authentication with X.509 certificates: Issue unique certificates to each device and encrypt communication with AWS IoT Core
- Per-device authentication: Prevent access from unauthorized devices
Data Encryption
- Encryption in transit: Encrypt all communication using TLS 1.2 or higher
- Encryption at rest: Enable server-side encryption (AES-256) for DynamoDB and S3
Access Control
- Principle of least privilege with IAM policies: Grant only minimum necessary permissions to each AWS service
- Family data isolation: Separate data by family using DynamoDB partition keys, preventing access to other families' data
Privacy Protection
- No voice data transmission: Voice is converted to text on device, voice data itself is not sent to cloud
- Data retention period: Achievement records stored indefinitely, cache data automatically deleted after 7 days (S3 lifecycle policy)
Demo
Explaining MUSUBiAi's operation flow using an actual device example.
MUSUBiAi Device
Photo of actual device (with component descriptions)
This device is a prototype created as an MVP. Developed based on Raspberry Pi 4B for basic function verification. The final device design is described in the "Future Challenges" section of "What I Learned."
Main device components:
- Achievement recording start button (yellow)
- Achievement recording stop button (black)
- LED indicator (green: standby, yellow: recording, yellow blinking: converting, red: error)
- Microphone (voice input)
1. Child Achieves Something
Moment when child achieves something (example: got 100 points on test)
2. Press Achievement Recording Button
Press yellow achievement recording start button
LED lights up and recording starts.
3. Speak Achievement Content
Child speaking achievement content (illustrated)
"Today, I got 100 points on my math test! I solved the difficult problems too!"
4. Press Achievement Recording Stop Button
Press black achievement recording stop button
Recording completes and LED changes to yellow blinking state. During this time, voice is converted to text on the device. When conversion completes, LED returns to green (standby state).
5. System Integration
Achievement record sent to MUSUBiAi system when online
When network is connected, achievement record is automatically sent to cloud.
6. Parent Receives Slack Notification (MVP)
Example of notification delivered to parent's Slack
System Features
- Simple: Easy for children to use (2 buttons)
- Fast: Real-time notification to parents (within seconds)
- Reliable: Operates offline and auto-syncs on reconnection
- Privacy: Voice data processed on device, not sent to cloud
What I Learned
Development Speed and Kiro's Effect
What surprised me most in this project was that by utilizing Kiro (AI assistant), I was able to complete basic functions in 2 weeks while balancing work.
I'm an IoT beginner, rarely handle IoT devices at work, and had limited knowledge of Raspberry Pi GPIO pins. However, by implementing while asking Kiro hardware questions, I could cover a wide range from software to hardware.
Technical Learnings
Advantages of Edge Computing
- Privacy: Process voice data on device without sending to cloud
- Latency: Process immediately without worrying about network delay
- Cost: Reduce cloud voice processing API costs
IoT Device Reliability Design
- Offline support: Continue operation during network disconnection
- Queuing: Max 100 items, 24-hour buffer
- Auto-reconnection: Automatically sync when network recovers
Serverless Architecture Scalability
- Lambda: Auto-scaling to handle traffic increases
- DynamoDB: Flexible scaling with on-demand capacity
- Operational cost: Usage-based billing suppresses initial costs
AI Integration Using AWS Bedrock
- Context engineering: Reference child's age, achievement content, and summaries/profiles stored in S3 (Knowledge Base) to provide personalized praise
- RAG utilization: Leverage Amazon S3 Vectors to create low-cost vector datasets that can be immediately utilized by Bedrock Agent
- Cost optimization: Minimize consumed token count through context optimization and reduce API call costs
Importance of Visibility in Spec Development
I realized that storing software and hardware structure in repository while maintaining visibility makes it easier for both developers and AI to understand.
Especially ASCII art is a format that AI can easily perceive, is easy to construct, and is also easy for humans to understand. By describing architecture diagrams and data flow diagrams in ASCII art, Kiro accurately understood the system and proposed appropriate implementations.
Importance of Consistency in Spec Development
AI is not omnipotent and reads and understands only specific files, so it may not be able to grasp the overall domain.
As the system grows, I learned the need for repository structure that maintains consistency and makes it easy for AI to grasp the whole. Maintaining consistency between documentation and code is key to successful AI-assisted development.
Specifically:
- Consistency of requirements definition, design, and implementation
- Clear reference relationships between files
- Unified naming conventions
Product Insights
Vision Validation
By actually creating a prototype and trying it myself, I gained a sense that the goals set in My Vision could be achieved.
I confirmed that the flow actually works: children press buttons to record achievements, parents receive Slack notifications, and praise children while referring to AI-recommended praise.
User Feedback
When I shared the idea with 2 families, I received positive feedback.
The following features were particularly well-received:
- Praise recommendation feature: "Helpful when I don't know how to praise"
- Review feature: "It's good to be able to look back on what children achieved"
Future Challenges
Phase 3: User Feedback Collection and AI Feature Expansion
In Phase 3, we plan to implement the following features:
- Chronicler Agent (Summary Generation AI): Automatically generate daily, weekly, and monthly summaries to visualize children's growth chronologically
- Profiler Agent (Profile Analysis AI): Automatically analyze children's interests, strengths, and growth patterns from achievement history
- LED Feedback Pattern Improvement: Enhance device visual feedback
Also, on March 7, 2026, we plan to exhibit MUSUBiAi at the JAWS DAYS 2026 corporate booth and receive feedback from actual users. JAWS DAYS is the largest one-day event hosted by JAWS-UG (Japan AWS User Group), which has over 60 branches nationwide, and attracts a wide range of participants from AWS beginners to advanced users. Through this event exhibition, we will grasp the needs of more families and use it to improve the product.
Voice Recognition Accuracy and Privacy Trade-off
Currently, MUSUBiAi uses Vosk (open-source voice recognition engine) to convert voice to text on the device. This protects privacy to the maximum by not sending voice data to the cloud.
However, Vosk's voice recognition accuracy has room for improvement, especially in recognizing children's voices and complex expressions.
If recognition accuracy is insufficient in the future, we are considering the following options:
- Migration to AWS Transcribe: Provides high-accuracy voice recognition but requires sending voice data to cloud
- Hybrid Approach: Provide as user-selectable option (privacy-focused vs accuracy-focused)
Option Comparison:
Method | Accuracy | Privacy | Cost | Implementation Status |
|---|---|---|---|---|
Vosk (Current) | Medium | Highest (processed on device) | Low | ✅ Implemented |
AWS Transcribe | High | Medium (cloud processing, encrypted communication) | Medium | 📋 Under consideration |
Ultimately, based on user feedback at JAWS DAYS 2026, we will decide considering the balance between accuracy and privacy.
Phase 4: Web Application and Final Device
In Phase 4, we will start developing the web application, providing a richer user experience with dashboards, analytics features, and achievement timelines. We will also leverage Bedrock services such as Bedrock Guardrails and Bedrock Flows to achieve flexible integration with the web application.
Also, the current prototype uses Raspberry Pi as is, so it's not in a state that can be carried by children. In Phase 4, we aim to evolve to a portable device with:
- Miniaturization (pocket size)
- Battery powered
- Improved durability
- Design easy for children to carry
Final device planned for Phase 4
Technical Challenges for Business Deployment
When considering business deployment as an actual product, we need to address the following technical challenges:
1. Device Registration Flow Improvement (JITP Implementation)
- Current challenges:
- Each device's certificate is included in OS image within OS image creation workflow
- This method cannot scale device numbers
- Individual image builds required for each device, increasing operational costs
- Solution:
- Implement device registration flow using JITP (Just-In-Time Provisioning)
- Mechanism where devices automatically register with AWS IoT Core on first connection
- Efficiently manage large numbers of devices through dynamic certificate generation and distribution
- Simplify manufacturing and distribution processes through device image standardization
2. Diversification of Notification Methods
- Current challenges:
- Slack integration is optimal for MVP development but cannot accommodate general users who don't use Slack
- Need to wait until Phase 4 web app
- Solution:
- Consider adding more common notification methods as Phase 3.5
- Email notifications (Amazon SES)
- Mobile push notifications (Amazon SNS Mobile Push)
- Fan-out architecture enables parallel implementation of multiple notification methods
3 . Cost Optimization
Current estimate (monthly cost per family):
Service | Usage | Cost |
|---|---|---|
Lambda | ~150 times/month (achievement record processing) | Free tier |
DynamoDB | ~150 reads/writes/month | Free tier |
S3 (Data) | ~10MB/month | $0.02 |
S3 Vectors (Knowledge Base) | Vector storage + queries | $0.05 |
Bedrock (Claude Sonnet 4.5) | ~150 times/month | $1.26 |
Bedrock (Titan Embeddings v2) | Knowledge Base embeddings | $0.02 |
IoT Core | ~150 messages/month | Free tier |
EventBridge | ~150 events/month | Free tier |
SSM Parameter Store | Slack token storage | Free tier |
Total | | ~$1.35/month |
When adding Chronicler/Profiler Agents in Phase 3:
- Additional cost: ~$0.40/month (summary generation and profile analysis)
- Total: ~$1.75/month/family
Optimization strategies at scale:
- Utilize Bedrock Provisioned Throughput (consider for 1000+ families)
- Strengthen cache strategy (reuse similar achievements)
- Optimize batch processing (simultaneous processing of multiple families)
Summary
Busy modern parents want to praise their children but lack specific information and praising skills. MUSUBiAi solves this challenge by capturing children's achievements in real-time and providing AI-powered specific praise recommendations.
With Kiro's support, even as an IoT beginner, I was able to implement basic functions in 2 weeks.
Going forward, we will complete Phase 3 and proceed with Phase 4 web app development to refine the product so we can provide value to more families.
Realize parents' desire to praise and nurture children's self-esteem.
That is MUSUBiAi's vision.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article