About Spokenly

Spokenly converts speech to text using cloud providers (OpenAI, Deepgram, Groq) via a Pro subscription or your own API keys, or local AI models (Whisper, Parakeet) that run entirely on-device. With BYOK, audio goes directly to the provider. With Pro, it is processed through Spokenly's backend. With local models, audio stays on your device.

Spokenly ships on macOS, Windows, Linux, and iOS. It includes AI text processing with modes and AI Instructions across its dictation apps. The macOS sideload build also includes an MCP server for AI coding agents and agent mode for voice-controlled automation. The iOS app adds system keyboard integration.

Spokenly gives you two ways to dictate. Cloud models (OpenAI, Deepgram, Groq) deliver the highest accuracy across languages and accents, powered by a Pro subscription or your own API keys. Local models (Whisper, Parakeet) run on-device for offline use and privacy, with model availability and hardware acceleration varying by platform.

On macOS, Windows, and Linux, press a shortcut, speak, and your text appears wherever the cursor is. On iOS, dictate through the Spokenly keyboard or app. No account is required for local models, and cloud transcription audio is not stored on Spokenly servers.

I'm Vadim Ahmerov, a former iOS developer turned indie maker. After years of building native apps for Apple platforms professionally, I went solo to create the tools I wanted to exist.

Spokenly started from a simple frustration: existing dictation tools were either too expensive, required cloud accounts, or did not respect privacy. I wanted an on-device option that supports 100+ languages and gives users full control over how their audio is processed.