Windows App & On Device AI fuels Speechify's dramatic growth with professionals and the enterprise. Privacy-first voice technology runs entirely on-device, with Copilot+ PCs (NPU from AMD, Intel and Qualcomm) and more Windows 11 PCs (GPU from Intel and AMD).
NEW YORK, March 31, 2026 /PRNewswire-PRWeb/ -- Speechify, the world's most used text-to-speech app, today announced the launch of its Windows application, bringing real-time text-to-speech and voice typing to Windows users with the option of fully on-device processing. Users can choose between cloud-based or on-device processing, and switch between them instantly at runtime.
The Leading Voice AI Productivity App on Windows
At Speechify, the mission has always been to make students and professionals more productive with Voice AI. With the Windows app, the team built real-time text-to-speech and speech-to-text that gives users the choice of cloud-based or fully on-device processing. When users choose on-device, their voice data never leaves the machine. The Windows ML stack and platform APIs made this possible across x64 and Arm64 architectures as well as chipsets across NPU and GPU, all from a single codebase.
"The Windows ML stack gave us a single runtime to deploy three production AI models on-device on x64 and Arm64 architectures along with NPUs and GPUs.) all through ONNX Runtime," said Raheel Kazi, Engineering Leader at Speechify.
Kazi continued, "And when we needed to ship on Snapdragon, the QNN execution provider put our transcription model on Qualcomm's NPU with FP16 precision without requiring us to start over. That's what accelerated our path to market: one inference foundation that scales from x64 to ARM64. Same holds true for addressing Intel and AMD chipsets."
NPU-Accelerated Transcription on Qualcomm Snapdragon
When Speechify came to Windows on Arm, the challenge wasn't whether on-device transcription was possible. It was whether it could feel real-time on a fanless Snapdragon laptop. ONNX Runtime's QNN execution provider provided the answer.
Speechify's speech-to-text model runs as a split encoder-decoder architecture routed directly to Qualcomm's Hexagon Tensor Processor in HTP burst mode with FP16 precision. The NPU handles the heavy inference while the CPU stays free for the rest of the app. At startup, Speechify's dependency injection container detects the processor architecture and resolves the right engine automatically. Users can also switch to cloud-based transcription if they prefer, and the app hot-swaps engines at runtime. Either way, they press a hotkey and get real-time transcription.
ONNX Runtime serves as the shared inference foundation: one runtime to optimize, one set of execution providers to configure.
A Real-Time Pipeline from Microphone to Output
On-device AI is only as good as the pipeline around it. Speechify engineered every step for seamless, gap-free performance. The Windows platform stack (NAudio for audio I/O, ONNX Runtime for inference, CsWin32 for native API access) provided the building blocks to make the experience feel seamless.
Built for Windows, Not Ported to It
Speechify on Windows is a WinUI3 application with deep native integration through CsWin32's source-generated P/Invoke bindings for over 30 Win32 APIs:
- RegisterHotKey gives system-wide shortcuts that work no matter what application has focus.
- SendInput auto-pastes transcribed text directly into whatever field the user was typing in.
- GDI screen capture powers OCR functionality.
- DwmSetWindowAttribute keeps the floating microphone indicator always on top with proper DPI scaling.
- Windows DPAPI encrypts authentication tokens at rest.
- A custom speechify:// protocol handler enables OAuth callbacks through the MSIX package manifest.
These are platform capabilities that make the difference between an app that runs on Windows and an app that belongs on Windows.
"Over a billion people on this planet use Windows," said Cliff Weitzman, Founder & CEO of Speechify. "With this Windows launch, we're making sure that reading, and now writing, is never a barrier, no matter what device you use or how you prefer to work. We're especially excited the opportunity in the enterprise given how many professionals have asked for Speechify on their PCs."
The Speechify Windows app is available now for x64 and Arm64 devices – exclusively from the Microsoft Store.
About Speechify: Speechify is the most used Voice AI productivity app in the world for anyone to have anything read out loud to them, including books, websites, PDFs, emails, school assignments, contracts, and docs. Speechify also offers voice typing, letting users dictate text in real-time with on-device or cloud-based transcription. Speechify's CEO & Founder, Cliff Weitzman, first launched Speechify's text to speech app to overcome his own dyslexia as a student at Brown University.
Media Contact
Rohan P., Speechify, 1 (747) 302-4454, [email protected], speechify.com
SOURCE Speechify
Share this article