Overview
Hush is an open-source noise suppression model developed by weya AI. It removes background noise, competing voices, and audio interference from real-time calls, making it ideal for voice AI agents and ASR systems. The model is lightweight (8 MB), runs on CPU in under 1 ms per 10 ms frame, and is ranked #5 on Hugging Face's Audio-to-Audio leaderboard.
Dayflow is a local-first macOS app that uses screen recording and AI to automatically journal your workday. It captures every activity (coding, meetings, research) and provides summaries for standups, 1:1s, and performance reviews. It is open-source under MIT license and supports both local and cloud AI models.
Feature Comparison
| Feature | Hush | Dayflow |
|---|---|---|
| Primary Function | Real-time noise suppression for voice AI calls | Automatic work journaling via screen recording and AI |
| Target Users | Voice AI developers, call centers, BFSI teams | Developers, freelancers, knowledge workers |
| Deployment | Cloud or data center (model size 8 MB) | Local macOS app (native SwiftUI) |
| Processing | Real-time on CPU (<1 ms per 10 ms frame) | Lightweight recording (~100 MB RAM, <1% CPU); AI analysis every 15 min |
| AI Model | Open-source noise suppression model (Hush v1.0) | Supports local (Ollama/LM Studio) or cloud (Gemini, ChatGPT, Claude) |
| Privacy | Deployable on own infrastructure; open-source | Local-first; all data stored on Mac; open-source (MIT) |
| Open Source | Yes (open-source model) | Yes (MIT licensed) |
| Key Use Case | Cleaning noisy call audio for ASR and voice agents | Automatically logging work activities for standups and reviews |
Pricing
Hush: The model is free and open-source. weya AI offers a 2-week pilot for enterprise implementation, where they help integrate and measure lift before full rollout.
Dayflow: The app is free and open-source (MIT). Using local AI models (Ollama/LM Studio) is free. Cloud AI (Gemini, ChatGPT, Claude) requires their respective paid subscriptions for API access. The app itself has no additional cost.
Pros and Cons
Hush
Pros:
- Real-time noise suppression on CPU, no GPU needed
- Lightweight model (8 MB) easy to deploy
- Open-source and auditable
- Reduces ASR errors in noisy environments
- Proven on Hugging Face leaderboard
Cons:
- Narrow focus: only audio noise suppression for voice AI
- Requires integration into existing voice stack
- Not a standalone product for end users
Dayflow
Pros:
- Fully automated work journaling with no manual input
- Privacy-first with local storage and local AI options
- Open-source (MIT) and transparent
- Context-aware activity recognition (e.g., distinguishes research from distraction)
- One-click summaries for standups and reviews
Cons:
- Requires macOS 13.0+; no Windows/Linux support
- Screen recording may raise privacy concerns in some workplaces
- Local AI can be GPU-intensive and drain battery
- Cloud AI accuracy depends on paid subscriptions
Verdict
Hush and Dayflow serve completely different purposes. Hush is a specialized noise suppression model for voice AI systems, ideal for developers needing clean audio in real-time calls. Dayflow is a personal productivity tool for macOS users who want automated work logging without manual effort. Choose Hush if you're building or improving voice AI call quality; choose Dayflow if you want to track and reflect on your daily work automatically.
Dayflow
Hush