This page is an AI-assisted translation of the original Japanese version.
In the event of any discrepancy between the English and Japanese versions, the Japanese version shall prevail.
Project Launch
Co-creation with Claude begins. Conceiving six AI systems and taking the first step in development.
Starting Situation
- Zero programming experience
- 25 years of theater sound support experience
- 15 years of international business experience
- A new challenge leveraging AI tools
Multilingual Meeting Concept Takes Shape
The blueprint for the top-priority system becomes clearer. Requirements for functioning in multilingual, multi-location environments are defined.
AI Multilingual Meeting Architecture Design v1.0 Complete
Architecture design for the top-priority system completed (approx. 86KB, 14 sections). The architectural framework designed for real-world operations is established.
Key Achievements
- Functional requirements defined (99-language speech recognition, 419-language translation)
- Technology stack selected (high-accuracy speech recognition engine, large-scale multilingual translation engine, etc.)
- Three delivery formats confirmed (software only, bundled with 1TB SSD, bundled with 2TB SSD)
- Pricing set (¥25,000)
Second AI System Concept Initiated
Design of the multi-camera video editing AI system begins. Workflow refinement and organization of complex requirements progressed.
Exploration for Design Quality Improvement
Collaborating with Copilot to explore architectural improvements. Third-party technical verification conducted.
License Investigation for Six Systems Complete
License investigation for all six systems including AI Multilingual Meeting completed. Commercially viable library selection and risk mitigation strategies confirmed.
Investigation Results
- Translation engine replaced (commercially viable through combination of OSS-compliant licenses)
- Commercial restrictions on image recognition models confirmed → alternative models considered
- CC-BY-NC status of voice generation models confirmed → commercial versions considered
- Commercially viable configurations confirmed for all systems
AI Multilingual Meeting v2.0 Multi-AI Integration Implementation
Multi-AI integration system design completed with multiple AI models. An innovative approach balancing accuracy and transparency established.
Technical Innovation
- Complementary system using multiple models of high-accuracy speech recognition engine
- Memory optimization through sequential processing (compatible with Core i3 8GB)
- Transparency ensured through visualization of inter-model differences
- EU AI Act compliant design (auditability and explainability)
Two-Week Architecture Sprint Complete
Over 13 days from 2025/12/30 to 2026/01/12, design work for six systems was initiated, and the architecture design for one system was completed (approx. 20,000 lines).
Key Achievements
- AI Multilingual Meeting Tool Architecture Design v3.0 complete
- Architecture concept completed for all six systems
- License investigation completed (all six systems)
- Development roadmap confirmed
AI English Meeting Pro Product Plan Complete
As a sibling product of AI Multilingual Meeting, the product plan for a Multi-AI integration speech recognition system specialized for English is complete.
Key Achievements
- Multi-AI coordination processing (high-accuracy processing combining multiple speech recognition AIs)
- Product plan document v1.1 complete (58KB)
- Market opportunity analysis (40,000 domestic organizations, 1,800,000 international companies)
- Pricing set (Personal edition ¥35,000, Enterprise edition ¥5,000,000)
AI Development and Testing Environment Acquired
Full-scale AI development environment secured. Hardware foundation for implementation phase development is in place.
An environment capable of implementing and testing large-scale AI models is now in place.
Key Specifications
- 64GB unified memory (compatible with large-scale AI models)
- High-speed GPU (MPS-compatible, optimized for parallel processing)
- Large-capacity storage (for model and dataset storage)
- Implementation phase development ready
Soul Resonant Works Official Website Officially Launched 🎉
https://sr-works.net/ is officially launched. Main site plus three legal pages are now live.
Foundation for subsidy applications and crowdfunding is complete.
The legal pages, prepared by a solo developer, achieved a Copilot rating of 9.55/10.
Published Content
- Top page published (seven AI systems, technology stack, CF information)
- Copilot rating 9.55/10 (BEM compliance, WCAG AA conformance, intellectual property protection design)
- Privacy Policy (covering major provisions of Japan's Act on the Protection of Personal Information)
- Terms of Service (standard SaaS terms, anti-social forces exclusion clause)
- Specified Commercial Transactions Act page (maintaining integrity with "under development" notation)
- Cross-browser verification completed (Chrome, Safari)
English Site Launched 🌐 — Now Bilingual
sr-works.net is now available in both Japanese and English. Every page has an English edition (/en/),
so visitors from abroad can read it too — the foundation for publishing in English is in place.
Published Content
- Eight new English pages published (/en/)
- AI-translation disclaimer added below the header on each page
- Japanese ⇄ English language toggle on every page
Development Blog Started ✍️
We started a blog about the background and story behind what we build (Japanese on note, English at /blog-en/).
It covers why we build these things and how they come together.
Published Content
- First Japanese post published on note ("Why I Started Soul Resonant Works — 25 Years in Theater, Zero Programming Experience")
- English blog (/blog-en/) opened, with its first post published on 4/22
- Updated roughly weekly since then, in both languages
Site Restructured — New "Everyday Tools" Track 🧰
As our lineup grows, we redesigned the site and added a new track for practical, local-first,
privacy-minded tools. We also refreshed the brand (logo, favicon, OGP).
Published Content
- New "Everyday Tools" track (first entry: CubePlot)
- Moved to a product-folder structure with clean URLs (/en/<product>/)
- Standardized on the non-www domain (sr-works.net)
- Refreshed brand assets (logo / favicon / OGP)
CubePlot Released 🎉 — SRW's First Product to Market
We released CubePlot, a tool that turns a CSV into a 3D scatter plot you can rotate and zoom.
No install, no sign-up — it's a single HTML file that runs in your browser and works offline.
CubePlot itself does not send the data you load outside your computer, and external network
traffic is blocked at the browser level (CSP). It's the first product Soul Resonant Works is bringing to market.
Published Content
- Product pages live (Japanese /ja/cubeplot/; English /en/cubeplot/ on 7/20)
- Free version: 3D view on the X/Y/Z axes, labels, slicing, trimming, and PNG/CSV export — no time limit (a small "CubePlot Free" marker on screen; exported PNGs carry a watermark + QR)
- Paid versions (commercial license / Student): color, color-scale, and size mapping (up to five parameters), rotation-video (MP4) export, watermark-free output
- Distribution: BOOTH (JP ¥4,980 incl. tax) / Gumroad (USD $39, live 7/20) / free Student version on request by email
- CubePlot itself sends no data externally — CSP blocks external network traffic, it runs offline, and the verify page explains how to check for yourself (/en/cubeplot/verify.html)