VoiSistant

VoiSistant splits its features across five modes, and you switch between them depending on the job. Dictation handles straight voice to text with instant paste. Chat is a conversational AI assistant with voice input. Consultant works as a meeting assistant with real-time translation. Transcribe converts audio files to text with speaker detection, and Quick Reply generates contextual responses from a screenshot of a conversation. Having them in one app means you're not juggling separate tools for talking, translating, and transcribing.

The raw transcript is only the starting point. VoiSistant sends it through a language model to fix grammar, adjust tone, and apply formatting before the text lands in your field. You connect whichever provider you prefer, including OpenAI, OpenRouter, Gemini, LM Studio, and Ollama. Why does that matter? If you already pay for one API, or want to run everything locally, you're not forced into someone else's stack. The trade-off is that the polish depends on the model you choose, so a weaker or rate-limited provider shows up directly in your output.

VoiSistant doubles as a voice translator. Translation runs inside the dictation flow rather than in a separate tab. You speak, the app converts and translates, and the result pastes where you need it, covering languages like English, Italian, Spanish, and Russian among others. Text to speech works the other direction, reading text aloud through Microsoft, Google Gemini, or OpenAI voices. Pick your voice engine and match the output to the situation.

The app stays out of your way by living in the menu bar, with a compact interface you summon rather than a window you keep open. Global hotkeys let you start and stop recording from any app. That's the part that makes dictation practical. You don't break your train of thought by switching windows, and the auto-paste step means the text appears right where you were already typing.

VoiSistant is built so you can keep processing on your own machine if you want to. Text to speech can run locally, and connecting a local LLM through Ollama or LM Studio keeps the cleanup step off the cloud too. Network access is only used when you enable a cloud provider, at least according to the app's own description. For people dictating sensitive material, that's a meaningful option rather than an all-or-nothing cloud setup.

After the AI step, VoiSistant can save the result to history, send it to email, run extra text cleanup, and export it to a file. It's a small pipeline, but it means the dictation output doesn't just vanish once pasted. If you dictate meeting notes or a task list, keeping a searchable history saves you from re-recording the same thing twice.

Transcribe mode turns existing recordings into searchable text. Point it at an audio file or a podcast episode and VoiSistant handles the AI transcription, with speaker detection to separate voices in the output. That covers the work you didn't do live, from interviews to conference talks. Nothing extra to install. It's all inside the same app.