Dicty.io

Offline voice dictation & docs librarian for developers

Productivity

Developer Tools

Artificial Intelligence

Visit WebsiteSee on Product HuntTwitter ⧉

Hi Product Hunt, I built Dicty because I talk out loud when I code, but tools like Superwhisper and MacWhisper always felt limited to just dumping raw text wherever my cursor was. As a developer, I wanted something that actually understood what I was working on: what branch I was on, what code I just changed, and where our technical notes live. Here is what Dicty does: 1. Truly local and private: Speech-to-text runs entirely on your machine using local Whisper models (Metal on Mac, NVIDIA/AVX2 on Windows). Your voice is never sent to a cloud server. 2. Connects to LM Studio: If you want AI to clean up your notes, format technical specs, or generate commit messages, you can point Dicty to LM Studio or Ollama and run the entire pipeline offline on your own GPU. If you prefer cloud models, you can also bring your own API keys. 3. Live translation and file transcription: You can speak in one language and have it transcribed into another on the fly. It can also record meetings and lectures (both microphone and system audio) or transcribe pre-recorded audio and video files offline. 4. Git and documentation sync: Instead of losing the reasoning behind your code changes, Dicty can take your spoken thoughts, look at your recent Git commits or diffs, and write structured updates directly into Obsidian, Confluence, or PR notes. 5. Native desktop builds: Native Swift on macOS and a lightweight Tauri/Rust app on Windows. It also includes local neural text-to-speech (Kokoro and Silero) for reading text aloud. I would really appreciate your honest thoughts: - If you already use Superwhisper, MacWhisper, or similar tools, what would actually make you consider switching to an alternative? - Does having local LLM support (via LM Studio) and Git context solve a real pain for you, or is simple voice-to-text enough for your daily work? Thanks for taking a look!