Typort

What It Does

No configuration. No accounts. Just power.

Ctrl + Space

Voice Mode

Hold, speak, release. Text appears at your cursor. Trained on your exact voice. Works anywhere.

Ctrl + Shift + Space

Compose Mode

Fact-check, rewrite, enhance your text. AI-powered editing without leaving your app.

Meeting

Meeting Mode

Real-time transcription with live AI insights. The co-worker who remembers every conversation.

Vault

AI Vault

Records every AI conversation you have. Searchable forever. Your future agents and AIs train on it.

Your Voice

Personal Model

A custom Whisper model trained on YOUR voice. Portable. Take it anywhere. Yours forever.

Coworker

AI Coworker

Surfaces insights from across your past work. Planning something? It finds that you discussed it months ago and what you decided.

Chrome

Chrome Extension

Captures every AI conversation from ChatGPT, Claude, Gemini, Perplexity, and Grok. Search across all of them in one place. Free.

Coming Soon

Slack Integration

Ask @typort in Slack and search your entire AI conversation history. Your team's collective AI knowledge, instantly accessible.

What You Get

Everything you need to get ahead of your competitors.

GPU Accelerated

Runs Whisper AI locally on your NVIDIA GPU via CUDA. Real-time transcription, not batch processing.

Works Offline

No internet required after setup. Your voice data stays on your machine. Always.

Types Anywhere

Simulates keyboard input. Works in any application that accepts text - even games, terminals, legacy software.

Learns Your Words

Correction dictionary adapts to your vocabulary. Technical terms, names, jargon - it remembers.

Multiple Languages

Whisper supports 99 languages. Switch in settings, no extra downloads.

System Tray

Runs quietly in your system tray. No window clutter. Right-click to configure.

No GPU? No Worries

Bring your own OpenAI key or use our cloud service. Personal model training, instant fallback, zero hardware needed.

Live Meeting Intelligence

Real-time transcription with AI insights delivered as you talk. Like a co-worker who remembers every conversation you've ever had — and pipes up at exactly the right moment.

Voice Commands

Corrections, screenshots, app switching — all by voice.

Not Another Electron App

Built on Rust and Tauri. Actually lightweight.

App	Size	Framework	Local	RAM
Typort	18 MB	Rust / Tauri	Yes	~50 MB
Misc. Electron Apps	400-800 MB	Electron	Yes	~300 MB
Otter.ai	Cloud	Web	No	N/A
Dragon	8+ GB	Native	Yes	~500 MB

Typort
                        Size: 18 MB
                        RAM: ~50 MB
                        Framework: Rust / Tauri
                        Local: Yes
                    

Misc. Electron Apps

Size: 400-800 MB RAM: ~300 MB Framework: Electron Local: Yes

Otter.ai

Size: Cloud only RAM: N/A Framework: Web Local: No

Dragon

Size: 8+ GB RAM: ~500 MB Framework: Native Local: Yes

Why Rust? What's the big deal?

Rust is the future of systems programming. It's not hype - the biggest tech companies on Earth are betting on it:

Microsoft is rewriting core Windows components in Rust for security
Linux kernel now officially accepts Rust code (first new language in 30+ years)
Google uses Rust in Android and Chrome
Amazon/AWS built critical infrastructure like Firecracker in Rust
Discord, Cloudflare, Dropbox - all running Rust in production

Why the shift? Rust is as fast as C/C++ but eliminates entire categories of bugs:

No null pointer crashes
No buffer overflows (the cause of most security vulnerabilities)
No data races in concurrent code
Memory-safe without a garbage collector

The type system catches bugs at compile time that would crash C++ programs at runtime. Microsoft found that 70% of their security vulnerabilities were memory safety issues - exactly what Rust prevents.

Typort isn't built on Rust because it's trendy. It's built on Rust because we want software that's fast, small, and doesn't crash.

Typort Vault

AI Conversation Governance. Every platform. Every prompt. Every response.

Your team uses 8 different AI tools. Where does that data go?

Claude. ChatGPT. Gemini. Perplexity. Copilot. Grok. DeepSeek. Your people are typing proprietary information into all of them, every day. None of it is centralized. None of it is searchable. None of it is auditable.

Captures every AI conversation your team has. Any platform. Any app. No extensions. No permissions. No setup.

"Works with every AI tool your team uses. No setup. No integration."

🔍

Full-Text Search

"What did marketing ask ChatGPT about pricing last month?" — answered in milliseconds. SQLite + FTS5.

📋

Audit Trail

Who asked what, when, on which platform. Full compliance reporting for legal, HR, and security.

🛡️

PII Detection

Flag when employees paste SSNs, customer records, or trade secrets into AI tools.

⚙️

Policy Enforcement

Define what's allowed. Block or warn when conversations violate your AI usage policy.

🏢

Air-Gap Compatible

Runs entirely on-premise. Zero cloud dependency. Your security team will actually approve this.

🤖

Agent-Ready

Vector-indexed knowledge base. When you deploy AI agents, they start from months of captured context.

Personal AI Model Training

Nobody else offers this. Not Wispr Flow. Not Dragon. Not Apple. Not Google.

Say "correct that to [word]" and Typort fine-tunes an actual Whisper neural network on your speech patterns. Not a dictionary — a real model trained on how you talk.

🌏

Accents

Australian, Scottish, Indian, Southern US, Korean-English — your personal model handles them all.

🗣️

Speech Impediments

Stutters, lisps, aphasia, dysarthria. Standard Whisper: ~60%. Trained model: 95%+.

🔬

Technical Jargon

Your industry terms, product names, acronyms. Say it once, correct it once, learned forever.

♿

Disability Access

100% free for verified users with disabilities. Local AND cloud. Forever. No expiration.

Your trained model is portable. It's a file. Take it to any machine. It's yours even if Typort disappears tomorrow.