Open Source AI Models 395

A browser to install, run, and control any application automatically, especially AI apps. Pinokio is a browser that lets you install, run, and programmatically control ANY application, automatically. It allows you to install, run, and manage any server application locally. You can download verified scripts from verified publishers. Pinokio supports various AI applications, including text-to-speech models, video generation tools, image editing systems, and more.

Open-source AI coding assistant for VS Code, enhancing coding efficiency and automation. Kilo Code is an open-source AI agent VS Code extension designed to enhance coding efficiency through code generation, task automation, and intelligent suggestions. It combines features from Cline and Roo Code, offering functionalities like orchestrator, architect, code, and debug modes. It supports various tools and integrations, customizable agent personas, and multilingual capabilities, aiming to be a comprehensive AI coding assistant.

Kommu offers AI-powered Level 2 autonomous driving solutions, making self-driving accessible. Kommu is a company revolutionizing self-driving technology with its KommuAssist 1s, an AI-powered Level 2 autonomous driving solution. It aims to make self-driving vehicles accessible to everyone. KommuAssist 1s offers features like Lane Keep Control, Adaptive Cruise Control, and Stop and Go functionality. It is driven by bukapilot, an open-source driving agent trained on Malaysian road data to provide a smoother and safer driving experience in Malaysia.

Native app for local AI model experimentation without complex setup or GPU. Local AI Playground is a native application designed to simplify the process of experimenting with AI models locally. It allows users to download and run inference servers without needing a full-blown ML stack or a GPU. The application supports CPU inferencing and model management, making AI experimentation accessible and private.

Open-source AI assistant for enterprise with RAG and LLM. MaxKB is an open-source AI assistant for enterprise, based on Large Language Models (LLM) and Retrieval-Augmented Generation (RAG). It integrates RAG pipelines, supports robust workflows, and provides MCP tool-use capabilities. It is widely applied in scenarios such as intelligent customer service, corporate internal knowledge bases, academic research, and education.

Making reasoning AI accessible through reinforcement learning for smart and safe machines. General Reasoning is dedicated to making state-of-the-art reasoning AI more accessible. They focus on developing machines that are smart and safe through reinforcement learning, aiming to help humanity. Their work explores the frontier of machine reasoning, with a focus on scaling RL compute and open-sourcing data and models.

Powerful 671B parameter MoE language model with state-of-the-art performance. DeepSeek v3 is a powerful 671B parameter Mixture-of-Experts (MoE) language model that offers groundbreaking performance. It is an AI-driven LLM with 671B total parameters (37B activated per token) and supports API access, an online demo, and research papers. Pre-trained on 14.8 trillion high-quality tokens, DeepSeek v3 delivers state-of-the-art results across various benchmarks, including mathematics, coding, and multilingual tasks, while maintaining efficient inference. It features a 128K context window and incorporates Multi-Token Prediction for enhanced performance and acceleration.

Focu is a mindful productivity app with local AI for private task management. Focu is a mindful productivity app that helps users transform their relationship with work through AI-powered guidance, meaningful conversations, periodic check-ins, and focused work sessions. It uses a locally running AI that works privately and offline on your Mac, ensuring data never leaves your device.

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, 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.

AI research company specializing in video generation and extension with Magi-1. Sand.ai is an artificial intelligence research and product company focused on advancing AI to benefit everyone. Their flagship product, Magi-1, is an autoregressive video model with top-tier quality output, designed for AI video generation and extension.

AI model router that optimizes LLM selection for accuracy and cost efficiency. Not Diamond is the world's most powerful AI model router. It is an end-to-end multi-model framework that helps developers accelerate development and improve accuracy with data-driven multi-model AI. By leveraging the best model for every query, Not Diamond helps you outperform every individual LLM on accuracy by up to 25% while reducing costs up to 10x. It can help you take any evaluation data for any set of models over any set of inputs and create an optimal routing algorithm tailored to your application.

Specialized AI model hosting for open, proprietary, and custom stacks EmpirioLabs AI is a low-cost AI infrastructure platform for developers. We host open-source models on our own infrastructure, run optimized proprietary endpoints, and bring model APIs, a browser playground, Hosted Agents, GPU Cloud, and generation templates into one place. Browse production-ready text, image, video, audio, 3D, search, and reasoning models, compare live pricing, test models in the playground, or launch video workflows from ready-made templates. Hosted Agents start at $5/month, GPU Cloud is billed by the second, and select models are discounted up to 90%.

AI text-guided video editing for character and scene transformation. Lucy Edit by Decart is the first foundation model for text-guided video editing. It allows users to seamlessly edit videos by changing outfits, adding accessories, or transforming characters, all guided by a simple text prompt. The model ensures that motion, face, and identity are perfectly preserved while the scene adapts around them. It is available in two versions: Dev, an open-weight community model, and Pro, a more consistent, detailed, and reliable version for creators seeking sharper results. Both versions support 480p and 720p editing and are accessible via API, platform, Hugging Face, and ComfyUI.

Self-hosted local AI platform for text, images, and speech Lemonade is a self-hosted local AI platform for running text, image, and speech models on your own hardware. It provides a GUI, CLI, REST API, and SDKs so you can chat locally, build AI apps and agents, and integrate local inference into existing tools. It is designed to be private by default, transparent, and flexible across major desktop and server platforms.

Unified API to route, manage, and analyze LLM requests. LLM Gateway is a platform that allows users to route, manage, and analyze their Large Language Model (LLM) requests across multiple providers using a unified API interface. It is fully open source and compatible with the OpenAI API format for seamless integration.

Novice is a desktop AI assistant for secure, offline document analysis and text editing. Novice is a desktop app for productivity that helps users get work done 10x faster. It's a safe place where AI can analyze documents and work alongside users securely and efficiently. Novice is an AI text editor and a local AI assistant that runs on your own computer, supporting models like DeepSeek, Llama 3.2, Phi, and Qwen2.5. It's designed for professional needs, allowing users to run AI locally and completely offline, ensuring no cloud processing and no data leaving the device. It can process PDFs, websites, DOCX files, and text documents locally, transforming them into a private knowledge base.

AI-powered document generator for essays, articles, and complex documents. Synapsy Write is a web application that uses OpenAI's GPT Generative AI models to help users easily generate text content. It's an AI-powered document generator that can create essays, articles, and complex documents using automations and AI.

Platform to run AI models on GPUs via API, pay-per-second billing. AI Tools 99 is a platform that allows users to run AI models and workflows on GPUs via an API. It automatically scales based on traffic and bills users only for the runtime, preventing GPU overcharges. Users can run and fine-tune open-source models and only pay for the time the GPU is running, billed by the second. When there is no activity, it scales to zero, and there are no charges.

Open-source ChatGPT alternative for building specialized and general-purpose chatbots. OpenChatKit is the first open-source ChatGPT alternative, offering a robust open-source foundation to build both specialized and general-purpose chatbots for various applications. It includes an instruction-tuned large language model, customization recipes, an extensible retrieval system, and a moderation model. It's not just a model release but a complete open-source project with tools and processes for continuous improvement and community contributions.

Open-source software to chat with PDF files for easy research and understanding. ChatWithMedia is an open-source software designed to make interacting with PDF files easy by allowing users to chat with their media. It aims to simplify research and understanding by turning it into a conversational experience.

Open-source 4K AI video model featuring native audio synchronization and local deployment. LTX-2 AI is an open-source, production-ready AI video and audio generation model designed to create stunning 4K videos with synchronized audio. The model supports text-to-video and image-to-video generation, achieving high quality output up to 50 FPS and 20 seconds duration per clip. LTX-2 is released under the Apache 2.0 license, allowing for local deployment via ComfyUI or integration through an API, offering users complete control and customization over their video generation workflows.

High-quality image to 3D model conversion for objects and human bodies. SAM 3D is an online tool powered by Meta’s SAM 3D research models, designed to convert a single image into a high-quality 3D model of objects or human bodies in seconds. It generates high-fidelity 3D models from any photo, offering accurate object and human-body reconstruction with exceptional speed directly in the browser, without requiring setup or a GPU. The platform features two specialized architectures: SAM 3D Objects for scene-aware reconstruction of detailed 3D shapes, textures, and layouts from a single image, and SAM 3D Body for precise human digitization, generating accurate 3D human pose and shape estimations using the Meta Momentum Human Rig (MHR). It leverages a 'Human-in-the-Loop' data engine, trained on nearly 1 million physical world images, to achieve unmatched robustness in diverse real-world scenarios. SAM 3D supports standard 3D export formats and is fully open-source under the Apache 2.0 License.

Free, open-source AI image generator with turbo inference and bilingual support. Z-Image is a free, open-source AI image generator that utilizes a 6B parameter diffusion model with 8-step turbo inference and bilingual prompt support. It is designed to produce photorealistic output, running efficiently on 16GB consumer GPUs. Z-Image is free for commercial use, requires no login, and is positioned as a top choice for professional creators, offering features like rapid generation, deep understanding of Chinese culture, and superior text rendering.