Introducing the AityTech Portfolio: Five Products, One Philosophy
PrivateAI, Legacy Dragon, AgentKits, Yomite, and StudyKits — a look at the AityTech product lineup and the privacy-first, on-device philosophy that connects them.
One Team, Five Products
AityTech started as a manufacturing AI and factory DX studio, but the same engineering philosophy — build it lean, build it to run locally where it can, don’t ask for more data than the job requires — has grown into a small portfolio of standalone products. Here’s where things stand today.
PrivateAI — AI Tools. Your Device. Zero Uploads.
PrivateAI is a collection of 15+ browser-based AI tools — OCR, text-to-speech, speech-to-text, background removal, and translation — that run entirely in your browser, GPU-accelerated, with no data uploaded anywhere and no account required. It covers text-to-speech with 322+ voices across 142 languages, speech-to-text in 99 languages, and offline translation across 200 languages. If you’ve read our post on building privacy-first AI apps, PrivateAI is that philosophy applied to the browser instead of a native app.
Legacy Dragon — Parse Ten Legacy Languages in Microseconds
Legacy Dragon is a parser for the languages that still run the world’s legacy IT estate: COBOL, JCL, PL/I, VB6, VB.NET, PowerBuilder, Assembly, SQL/DB2, CICS, and REXX. It parses a 1,200-line COBOL program in about 6 milliseconds, builds interactive AST graphs for exploring program structure, handles Shift-JIS/EBCDIC/DBCS encodings, and ships as a single binary for locked-down environments. We wrote more about the thinking behind it in our legacy modernization guide.
AgentKits — Open-Source AI Agent Kits
AgentKits is our open-source, MIT-licensed toolkit of AI agent kits for Claude Code, Cursor, and GitHub Copilot — 35+ agents, 178+ commands, and 68+ skills covering marketing, sales, legal, and engineering workflows. It’s the most mature product in the portfolio and stays free and open source.
Yomite (読み手) — Offline Japanese Text-to-Speech
Yomite is an offline Japanese text-to-speech reader for iOS, iPadOS, and macOS. Feed it Japanese text and it reads it back to you — no internet connection required. It’s a focused tool for a specific, real need: reading Japanese text aloud without depending on a cloud service.
StudyKits — Certification Exam Prep, Without the Bloat
StudyKits is our portfolio of certification study and exam-prep apps, built around the same idea as everything else here: focused tools that do one job well instead of trying to be everything.
Why This Portfolio, Why Now
We didn’t set out to build five separate products. Each one started as a tool we needed ourselves — a way to transcribe something without uploading it, a way to make sense of a legacy codebase, a way to read Japanese text aloud on a train with no signal. Shipping them as standalone products, rather than keeping them internal, is our bet that other people run into the same problems.
The connective tissue is simple: process locally when you can, don’t collect data you don’t need, and ship something narrow and reliable rather than broad and mediocre. If any of these solve a problem you have, they’re all live and linked from our products page.
Questions about any of these, or want to compare notes on a similar build? Reach out at hello@aitytech.com.
See Our Work
From MinuteAI to AgentKits — explore the products and projects we've shipped.
View PortfolioRelated Articles
The Summary Doesn't Leave Either: Why MinuteAI Runs Meeting Notes Through a Local LLM
Most privacy-focused meeting recorders still send the transcript to a cloud LLM to generate the summary. MinuteAI's GGUF/llama.cpp pipeline keeps that step on-device too — here's why that gap matters more than it sounds.
Guides410 Million Reasons Not to Upload: PrivateAI and the Shadow AI Problem
ChatGPT alone triggered over 410 million DLP policy violations in 2025. Most of that leakage isn't malicious — it's employees doing normal work through tools that happen to have a server in the loop. Here's why PrivateAI's browser-only architecture sidesteps the problem structurally.
GuidesWho Said What: Why MinuteAI Ships Speaker Diarization as a Download, Not a Bundle
MinuteAI's speaker diarization runs on pyannote.audio and PyTorch — too heavy to bundle into a lightweight SwiftUI app. Here's why it ships as an optional 234MB download instead, and how that fits a wider shift toward on-device diarization.