Kekoso

Kekoso is a macOS app that turns speech into text entirely on your device, covering dictation, file and YouTube transcription, call recording, and a local MCP server for AI agents. It runs on Apple silicon with models like Parakeet, Whisper, and SenseVoice, supporting up to 99 languages. It is for professionals and developers who want privacy-focused, subscription-free speech-to-text without cloud uploads.

How it works

Kekoso captures audio on your Mac and processes it locally using the Neural Engine, with no network calls for recognition. You press a hotkey, speak, and the text appears in the field where your cursor is. The model stays warm in memory to avoid cold starts, and audio is discarded after recognition.

Key capabilities

Beyond dictation, Kekoso transcribes audio and video files (MP3, M4A, WAV, MP4, MOV, and more) with timestamps, exports as TXT, SRT, or VTT, and handles YouTube links for material you own. It records calls split into microphone and system tracks with speaker labels, and runs a local MCP server with twelve tools for Claude Code, Claude Desktop, and Codex.

Who it is for

Kekoso is for Mac users who dictate into apps, transcribe interviews or meetings, and developers who want to give AI agents transcription without cloud dependencies. It supports up to 99 languages for recognition, and the interface is available in ten languages including English, German, Spanish, French, Italian, Portuguese, Russian, Simplified Chinese, Japanese, and Korean.

Pricing and limits

Kekoso costs $29 once, with a 7-day free trial and no card required. It includes all features, no per-minute billing, use on up to three of your own Macs, free updates through the 1.x line, and a 30-day refund. It works offline indefinitely with no license check, and payment is handled by Lemon Squeezy.