Kenzy

Create a private, locally-hosted smart assistant with voice control and face recognition using this open-source Python framework for secure home automation.

Free personal assisting

About Kenzy

Kenzy is an open-source AI foundation designed to act as a local, private alternative to mainstream smart assistants like Siri or Alexa. Built entirely in Python, it allows users to create a personalized digital assistant that can see, hear, and respond to interactions without relying on cloud-based processing. Its primary purpose is to provide a secure environment for home or business automation where sensitive data never leaves the local network, addressing the growing privacy concerns inherent in proprietary solutions from major tech corporations. The platform operates on a modular, distributed architecture composed of specialized device containers that handle specific input and output tasks. These include the Listener for speech-to-text via OpenAI's Whisper model, the Watcher for face and object detection using OpenCV, and the Speaker for synthesized audio output. At the center of the system sits the Skill Manager, which acts as the brain to parse intents using neural networks. Because it uses pre-trained, optimized models, Kenzy functions entirely offline, utilizing TCP/IP sockets for communication between modules to allow the processing load to be spread across multiple machines on a local network. This tool is ideal for developers, privacy advocates, and DIY tech enthusiasts who want to build custom AI experiences without sacrificing data security. It is particularly well-suited for users in the smart home space or industrial environments where external internet connectivity is either restricted or undesirable. Since it is MIT Licensed, it is also business-friendly, enabling companies to integrate AI-driven human interaction into their own products or internal tools without the licensing complexities or recurring subscription costs associated with commercial cloud APIs. What sets Kenzy apart is its commitment to 100% standalone operation and its modular scalability. Unlike mainstream assistants that send voice snippets to the cloud for processing, Kenzy handles all inference locally, ensuring that private conversations and video feeds remain within the user's control. While it may trade off some extreme accuracy compared to massive cloud models, it provides a transparent, customizable, and lightweight framework that can run on modest hardware like an i5 desktop or distributed across hardware clusters.

Kenzy pros & cons

Pros

Operates 100% locally without requiring internet access for enhanced privacy.
Modular design allows distributing processing tasks across multiple networked devices.
MIT Licensed, making it suitable for both personal and commercial applications.
Includes face and object recognition capabilities alongside voice interaction.
Open-source Python foundation makes it highly customizable for developers.

Cons

Higher hardware requirements compared to cloud-based alternatives for smooth performance.
May have slightly lower accuracy than proprietary cloud models due to optimized local model sizes.
Mover module for servo/controller interaction is currently unavailable and listed as coming soon.
Requires technical knowledge of Python and networking to set up and customize.

Kenzy use cases

Privacy-conscious homeowners can build a custom voice assistant that manages smart devices without sending audio data to external servers.
Software developers can use the Python-based framework to prototype and deploy unique AI-human interaction experiences in local environments.
Businesses can integrate secure, offline face and object recognition into their physical infrastructure without recurring API costs.
Open-source contributors can expand the platform's capabilities by building and sharing custom skill modules within the community.
Hobbyists can set up a distributed hardware network using Raspberry Pis and older PCs to create a powerful home automation hub.

Kenzy features

speech recognition
speech synthesis
modular architecture
face/object detection
intent determination
offline processing
custom skill modules
distributed i/o support

Kenzy pricing

Is Kenzy free? Yes, Kenzy has a free plan.

Open Source

Free

Self-hosted environment
MIT Licensed
Offline processing
Speech recognition
Face & object detection
Modular architecture
Python-based
No internet required

Kenzy FAQs

Does Kenzy require an active internet connection to work?

No, Kenzy is designed to be a 100% standalone codebase that does not require any network or external connectivity. It uses pre-built, optimized machine learning models that are loaded directly into your local hardware for processing.

What kind of hardware is needed to run Kenzy?

While it can run on a Raspberry Pi, it is recommended to use something more powerful for optimal performance, such as an i5-4600 CPU with 8GB of RAM. The modular design also allows you to spread the processing load across multiple computers.

Which AI models does Kenzy use for its core functions?

Kenzy utilizes OpenAI’s Whisper model for speech-to-text, OpenCV for object detection, Festival for text-to-speech, and MyCroft AI’s Padatious for intent parsing. These models are selected for their balance between size and local performance.

Can I add my own custom features to the assistant?

Yes, Kenzy features a modular design that allows you to expand its functionality by building your own skill modules. You can integrate new hardware or create custom logic to fit your specific automation needs.

Is Kenzy suitable for commercial use?

Yes, Kenzy is released under the MIT License, which is business-friendly and allows for significant freedom in how the software is used and integrated into other projects.

Ratings & reviews

No reviews yet. Be the first to share how Kenzy worked for you.