On-device AI 20

Open-source browser AI components for private, offline applications LocalMode UI is an open-source, MIT-licensed toolkit and shadcn-style registry for building privacy-first AI interfaces that run entirely in the browser. It provides more than 100 copy-owned React components and 36 installable blocks for chat, RAG, vector search, vision, audio, agents, privacy, and developer tools. Models run locally through WebGPU or a WebAssembly fallback, so applications can work without servers or API keys and keep user data on the device.

On-device voice assistant for conversational AI coding SKI is an on-device voice assistant for AI coding agents such as Claude Code, Cursor, Codex, Gemini CLI, Windsurf, and OpenClaw. It lets developers speak naturally to their coding agent, have the agent build and manage code, and hear its responses aloud. Unlike dictation tools, SKI provides a full two-way voice conversation with local speech-to-text, neural voice synthesis, full-duplex interruption, agent status updates, multi-project support, and optional meeting participation. Voice data, transcripts, and local meeting recordings remain on the user's computer and are not uploaded.

Private AI chat, local agents, and image and video generation in one suite DreamFort is a private, all-in-one AI suite for confidential chat, AI image and video generation, and local AI agents. It provides access to models such as ChatGPT, Claude, Gemini, Grok, DeepSeek, Seedance, Veo, Kling, Wan, GPT Image, and Seedream through an identity-stripping privacy architecture. Users can chat anonymously, create images and videos with permissive models, and run code or analyze files locally in a browser sandbox. DreamFort uses local encryption, Oblivious HTTP relays, and trusted execution environments to separate user identity from AI requests. Chats and generated content are stored locally on the user's device, while usage is paid through credits that can be purchased by subscription or one-time top-up.

Offline AI app for iPhone, iPad, and Mac, ensuring total privacy and customizable AI models. Privacy AI is a mobile and desktop application that allows users to run advanced AI models offline, ensuring complete data privacy. It features models like LLaMA, Mistral, Phi3, StableLM, and Gemma2, with customizable settings for fine-tuning interactions. The app is available for iPhone, iPad, and Mac, offering both free and premium options, including a one-time purchase model.

No-code, on-device AI with voice and RAG for local document interaction. RAGDrive.com is a No-Code On-Device AI with Voice & RAG. It is a user-friendly local AI inferencing solution. It allows users to 'talk' to their documents using Retrieval Augmented Generation (RAG) technology. Users can choose to operate completely offline and private or leverage inference providers when they need more computing power.

Private AI tools for complex enterprises, focusing on financial, legal, and regulatory compliance. LLMWare AI provides pioneering AI tools built for financial, legal, compliance, and regulatory-intensive industries in private cloud environments. It offers an innovative end-to-end solution, from specialized LLMs to RAG frameworks, enabling complex enterprises to leverage AI securely and efficiently. LLMWare focuses on private AI solutions, allowing users to run AI models locally on PCs and laptops, ensuring data privacy and control. The platform supports features like on-device document search, natural language SQL queries, and more, all without requiring an internet connection after initial setup.

Chat with leading AI models on your device. InstaLM allows users to chat with leading AI models directly on their devices. It offers features such as enriched conversations, multi-bot management, diverse AI model access, advanced tool integration, and multilingual support.

Private Mac AI assistant for searching, remembering, and managing work locally. Zenpa is a private AI assistant for Mac that remembers work across emails, chats, notes, documents, and meetings. It provides answers with sources, maintains a local memory vault, drafts messages, transcribes meetings, and offers system-wide dictation. All processing and stored data remain on the user's Mac.

Privacy-focused macOS app for instant on-device dictation and AI-powered text processing. Spoke is a native macOS dictation application that provides instant voice-to-text transcription directly into any text field. It runs a high-performance local speech recognition model via CoreML, ensuring that audio never leaves the device for maximum privacy. Users can hold a customizable keyboard shortcut to speak and see their words appear instantly at the cursor. The app also features 'AI Skills,' allowing users to process transcriptions on the fly for tasks like translation, grammar correction, or summarization using their own API keys from providers like OpenAI, Anthropic, or local models via Ollama.

A personal system with on-device AI and storage for secure file management and private AI chat. PERSYS is a personal system with on-device AI and storage, designed to keep personal files safe and enable intimate conversations. It can be accessed via its onboard hotspot or connected to a home WiFi network.

Privacy-focused macOS menu bar app for fast, on-device translation without internet requirements. Plae is a macOS menu bar application designed for fast, private, and on-device translation. It allows users to translate text instantly using a global keyboard shortcut without leaving their current application. The app stands out by performing all translations locally on the machine, ensuring no data ever leaves the computer. It offers three distinct translation engines: the native Apple Translation, context-aware Apple Intelligence, and an independent built-in AI model based on Google's TranslateGemma. Key features include a full-screen presentation mode for sharing translations, translation history, and complete offline functionality.

On-device AI solution for mobile devices, reducing server costs and enhancing performance. ZETIC.MLange is an on-device AI solution using NPUs to run AI models directly on mobile devices. It supports various SoC NPUs, providing optimized AI models and easy implementation across platforms like Android, iOS, and Windows. ZETIC.ai helps build zero-cost on-device AI, enabling server-less AI and breaking free from GPU clouds by utilizing any NPU and any OS. It transforms AI with the world’s fastest on-device AI software, solving AI companies’ challenges with NPU-powered on-device AI solutions, delivering the fastest transformation and runtime performance.

Run any AI model locally - text, speech, vision understanding and image generation Nexa SDK is an on-device inference framework that runs any model on any device, across any backend. It runs on CPUs, GPUs, NPUs with backend support for CUDA, Metal, Vulkan, and Qualcomm NPU. It handles multiple input modalities including text 📝, image 🖼️, and audio 🎧. The SDK includes an OpenAI-compatible API server with support for JSON schema-based function calling and streaming. It supports model formats such as GGUF, MLX, Nexa AI's own .nexa format, enabling efficient quantized inference across diverse platforms.

Turns the MacBook notch into an interactive AI and utility control center. NotchIA is a native macOS application that transforms the MacBook notch or menu bar into an interactive control center and cockpit. It brings critical information right to eye level by integrating a variety of utilities including a multi-source media player with synced lyrics, scrollable calendar and reminders, system HUD replacements for volume and brightness, a productivity toolkit with clipboard history and file converters, and active tracking for live AI developer assistants like Claude Code, ChatGPT Codex, and GitHub Copilot. It also features a localized news brief powered on-device by Apple Intelligence from user-pasted RSS feeds.

PaperClip: An on-device AI tool for researchers to save and find research paper findings. PaperClip is a tool designed for AI researchers, particularly those in machine learning, computer vision, and natural language processing, to save and easily find research paper findings. It operates using on-device AI, eliminating the need for API calls or server costs, and is offered for free. It helps users keep track of daily AI paper reviews, memorize details from papers, and find back important findings from AI research papers, ML blog posts, and news.

Offline, private AI agent for document search & understanding. Hyperlink is a local-first AI agent that understands your files privately—PDFs, notes, transcripts, and more. No internet required, ensuring data stays secure, offline, and under your control. It functions as a personal AI agent designed to help you recall and create information using your personal data directly on your device. Hyperlink finds information across your local folders and surfaces key insights you'd otherwise miss, serving as a Glean alternative built for personal or regulated use.

Platform for on-device speech AI, enabling speech recognition and wake word detection. Wavify is a one-stop-shop for voice AI, providing a platform for on-device speech AI. Software engineers can embed features like speech recognition and wake word detection into any software. It offers SOTA models and a cross-platform inference engine, optimized for speed and privacy. Wavify supports multiple languages and runs on various platforms, including Linux, Mac, Windows, iOS, Android, Web, Raspberry Pi, and embedded systems.

MacOS app for email management and task automation with AI-powered workflows. Inbox AI is a MacOS application designed to manage email and automate everyday tasks using custom AI-powered workflows. It offers both cloud-based and privacy-first on-device AI options. It natively connects with Apple Reminders and integrates with other apps through API or file-based commands. Users can have conversations with AI, automate tasks, process emails, and capture information via voice. It allows users to build custom AI-powered assistants that can answer questions and perform actions, such as drafting emails, remembering information, discussing text, rewriting content, replying to messages, searching the web, opening apps, and transcribing text.

Test Chrome's built-in offline AI model, Gemini Nano. The Chrome Developer Edition has integrated Gemini Nano, enabling offline use of built-in AI models. This ensures seamless operation even without an internet connection, while also guaranteeing data security and fast response times. The page allows you to run Chrome built-in large language model AI locally in your browser and test Gemini Nano.

Private, on-device AI browser extension powered by Ollama, ensuring no cloud dependency. NativeMind is a private, on-device AI assistant that operates as a browser extension. It brings the latest AI models, powered by Ollama and leading open-weights models like DeepSeek, Qwen, Llama, Gemma, and Mistral, directly to your browser. The core value proposition is total privacy and no cloud dependency, ensuring all data and AI processing remain on your local device.