Otter.ai
AI meeting assistant for transcription and automated note-taking
WhisperWriter
Instantly transcribe speech to any active window with a keystroke
WhisperWriter captures microphone audio and instantly transcribes it into the active window using OpenAI Whisper, controllable via a single keyboard shortcut and fully configurable UI.
Overview
Overview
WhisperWriter is a lightweight desktop utility that listens for a configurable keyboard shortcut, records from the microphone, and writes the resulting transcription directly into the currently focused application. It targets writers, developers, researchers, and anyone who benefits from hands‑free text entry.
Capabilities & Deployment
The app offers four recording modes—including continuous, voice‑activity detection, press‑to‑toggle, and hold‑to‑record—allowing users to match their workflow. Transcription can run locally via the faster‑whisper package (with optional CUDA‑12 GPU acceleration) or through the OpenAI Whisper API, selectable in the settings window built with PyQt5. After installation via a Python virtual environment, users launch the program, configure shortcut keys, model preferences, and language options, then start the background listener. The status window can be hidden for a truly unobtrusive experience.
Getting Started
Prerequisites are Git and Python 3.11; GPU users must install cuBLAS and cuDNN for CUDA 12. Detailed installation steps are provided in the repository README, and the UI guides users through initial configuration without manual file edits.
Highlights
PyQt5 settings window with full UI configuration
Four flexible recording modes for different workflows
Local Whisper via faster‑whisper or remote OpenAI API
Customizable activation shortcut and language options
Pros
- Runs silently in the background, activated by a single shortcut
- GPU acceleration available for fast local transcription
- Supports both local models and OpenAI API for flexibility
- All settings adjustable through an intuitive UI
Considerations
- Requires Python 3.11 and optional NVIDIA libraries for GPU use
- Initial setup may be challenging for non‑technical users
- Limited to OpenAI Whisper models; no alternative engines
- Potential latency when using the remote API
Managed products teams compare with
When teams consider WhisperWriter, these hosted platforms usually appear on the same shortlist.
Looking for a hosted option? These are the services engineering teams benchmark against before choosing open source.
Fit guide
Great for
- Writers needing hands‑free dictation
- Developers who want quick code comments via voice
- Researchers transcribing interview audio on the fly
- Users who prefer keyboard shortcuts over mouse interaction
Not ideal when
- Environments without a Python runtime
- Systems lacking microphone access or permissions
- Scenarios requiring multi‑language models beyond Whisper's support
- Live‑stream captioning where ultra‑low latency is critical
How teams use it
Live note‑taking during meetings
Transcribed speech appears directly in your document editor, freeing you to focus on discussion.
Coding with voice commands
Write comments or code snippets by speaking, instantly inserted at the cursor position.
Research interview transcription
Capture interview answers without manual recording, then edit the text immediately.
Hands‑free email drafting
Dictate email content and have it appear in your mail client with a single shortcut.
Tech snapshot
Python100%
Frequently asked questions
Do I need an OpenAI API key to use WhisperWriter?
An API key is only required if you choose the remote OpenAI Whisper API; the default configuration uses a local model.
Can WhisperWriter run on Windows?
Yes. The application works on Windows, macOS, and Linux as long as Python 3.11 and the required dependencies are installed.
How can I change the activation shortcut?
Open the Settings window, locate the 'activation_key' option, and enter your preferred key combination.
What GPU drivers are needed for faster‑whisper?
CUDA 12 with cuBLAS and cuDNN 8 are required; compatible NVIDIA drivers must be installed.
Is there a way to hide the status window?
Yes, the status window can be disabled in the configuration options within the Settings UI.
Project at a glance
Dormant- Stars
- 999
- Watchers
- 999
- Forks
- 157
LicenseGPL-3.0
Repo age2 years old
Last commitlast year
Primary languagePython
Last synced 4 hours ago