saytotype.ai

What It Is

SayToType is a speech-to-text dictation app built around the idea of “press a hotkey, speak, and type.” It supports Windows, macOS, iOS, and Android. Its core experience is automatically inserting the transcription at the current cursor position, so it can be used in browsers, messaging apps, editors, and even terminals, reducing copy-paste and app switching.

Core Capabilities

The product supports fast transcription, 47 speech recognition languages, automatic translation, custom hotkeys, and custom modes. Modes can be configured for scenarios such as emails, casual messages, and code comments, with different languages, providers, output formats, and prompts. AI provider options are fairly flexible: by default, users can use SayToType’s cloud transcription; on macOS Apple Silicon, local Whisper models are supported for offline processing without uploading audio; Premium users can also connect OpenAI, Mistral, Deepgram, or OpenAI-compatible APIs.

Pricing and Privacy

The page only states that there is a free tier and a Premium plan, and that advanced local models and custom providers require Premium, but it does not disclose specific pricing. Cloud transcription is included in the free tier, while the local Standard and Standard English models are free. Using automatic translation or custom prompts increases minute consumption by about 10-20%. On privacy, the page states that data is not stored and that processing happens in real time; local mode is handled entirely on-device.

Pros and Cons

The strengths are smooth cross-app input, simple hotkey activation, support for multiple providers, and local offline models, making it suitable for embedding voice input into everyday workflows. The drawbacks are also clear: Windows does not currently support local transcription; a Chinese interface is not listed among the supported languages, and it is unclear whether Chinese speech recognition is available; pricing and payment methods are missing; and the text does not make it possible to judge the accessibility of the cloud service or third-party APIs from mainland China.

Who It’s For and Access from China

It is a good fit for users who frequently write emails, chats, notes, code comments, and similar content, as well as privacy-conscious users willing to use local models on a Mac. Users in China should specifically test connectivity to the official website, login, cloud transcription, and the OpenAI/Deepgram/Mistral APIs. If access is limited, alternatives include system dictation, local Whisper tools, MacWhisper, or domestic speech input products such as iFlytek-style solutions.