Local & Self-Hosted AI Tools

LunarGate is a self-hosted AI gateway for teams whose apps use multiple model providers. It gives those apps one OpenAI-compatible endpoint and keeps routing, retries and fallback rules in your infrastructure. The gateway is open source under the MIT license and runs on macOS and Linux, with a container image also available.

AudioNotes turns audio and video into transcripts and structured Markdown notes on your computer. It's for reviewing meetings, interviews, courses and voice memos while keeping the recordings and resulting notes locally.

Voice-Pro is a locally run speech processing and dubbing app for creators working with recordings or multilingual video. Its Gradio browser interface brings audio preparation, transcription and speech generation into one workspace. Internet access is required.

Hindsight is a memory system for developers whose AI agents need to carry facts and experience across sessions. It combines stored memories into observations and maintains answers to recurring questions as more information arrives.

CodeFox CLI reviews code changes for developers and teams who want feedback in the terminal or automated pull request checks. It considers the surrounding project code when assessing a git diff.

GLaDOS is a local AI voice assistant for people who want a conversational companion modeled on the personality core from Valve's Portal games. It combines a custom voice and dry humor with awareness of its surroundings.

Genkit is Google's framework for developers building AI agents and chatbots into web and mobile apps. It brings model connections, application workflows and testing tools into one framework.

F5-TTS is a local text-to-speech system for developers and creators generating speech on their own hardware. Its Gradio app provides a browser interface for speech generation, with tools for training and fine-tuning models.

miel is a self-hosted Gmail client for people who want AI help sorting mail while keeping the app and its database on their own machine or server. It ranks messages by priority and proposes labels, including new ones. You review suggestions before applying changes to Gmail.

OpenHuman is a personal AI assistant for people who want agents to act across their work apps and carry out repeatable tasks. Its orchestration engine manages multiple agents and workflows in one place.

XGBoost is a machine learning library for training and running gradient-boosted tree models on your own hardware. Data scientists use it for classification, regression and ranking, with parallel computation for larger workloads. The code is licensed under Apache 2.0.

SkyReels-V2 is the downloadable video model family for developers and researchers who want to generate footage on their own GPUs. It includes Python inference code and model weights, distinct from the SkyReels web playground. Generation runs on your hardware.

Mosec is a self-hosted model serving framework for developers turning trained machine learning models into APIs. It runs on Linux and macOS, on your own hardware or cloud servers, and is open source under the Apache 2.0 license. Its main focus is handling concurrent requests while making better use of CPU and GPU resources.

Paperless-AI is an extension for Paperless-ngx users who want to sort incoming documents automatically and ask questions about their archive. The project is no longer maintained.

CSBDeep is a Python toolbox for researchers who want to restore fluorescence microscopy images using content-aware restoration (CARE). It covers training restoration models and applying them to images, with deep learning built on Keras and TensorFlow.

Parakeet is NVIDIA's speech recognition model family. The linked English model, parakeet-tdt-0.6b-v2, is for developers building transcription services, subtitles and voice applications with a model they can run locally.

PaddleClas is a PaddlePaddle toolkit for developers and researchers building image classification and recognition applications on their own hardware. It covers fixed-category classifiers and image matching for large or frequently changing catalogs.

PaliGemma is a family of vision-language models for developers and researchers building applications that interpret images alongside text prompts. Pretrained models need fine-tuning to produce useful results; mix models can handle common tasks without further tuning.

Dyad is a desktop AI app builder for people who want to create web apps from written descriptions while keeping access to the code. It keeps project files on your machine and can work offline when you use local models.

Codex Orchestration is a plugin for developers who want different models to plan, review, design and implement a coding task. It runs locally, but its Claude and Kimi K3 routes require authentication and send task content to cloud services.

Wan2GP generates media on your own computer for creators who want to work with several AI models in one browser interface. It builds on Wan-Video/Wan2.1 and focuses on making generation accessible on GPUs with limited memory.

Codel is a self-hosted AI agent for people who want software to carry out multi-step tasks and projects using a terminal, web browser, and text editor. It chooses and performs the next action itself, with task execution inside a sandboxed Docker environment. You access its interface through your browser.

Lumina-Image 2.0 generates images from text on your own hardware. It's for developers, researchers and people who want a model they can use in visual workflows or adapt with their own image collections.

Hugging Face CLI is a terminal tool for developers finding models and datasets for local AI projects or publishing their own work. It connects to the Hugging Face Hub, where those resources are hosted.

DeepEthogram assigns predefined behavior labels to individual video frames. It's designed for neuroscience researchers measuring animal activity, and can also serve other projects that need action detection at frame level.

Magenta RealTime is a local AI music model and synthesis engine for musicians playing AI instruments and developers building interactive music apps. It generates a continuous audio stream on supported Apple Silicon Macs.

Outlines is a Python library for developers who need LLM responses in a defined format. It constrains generation to produce structured data for tasks such as extracting service tickets from emails or limiting an answer to allowed choices.

Discord Ollama Bot lets a Discord community chat with language models hosted on the owner's hardware. It's for server owners who want to choose their own models and make an AI assistant available inside Discord.

Anti Slop is a set of skills for developers who want AI coding agents to avoid generic interfaces and filler copy. It checks the agent's work while leaving visual direction to you.

Vibe-Trading is an installed Python research app with terminal and web interfaces for traders and analysts testing strategies or reviewing trades. The workspace runs on your machine, but its AI requests go to OpenRouter by default. Research also fetches external market data and web content.

Vanna is a Python framework for adding database chat to analytics applications and internal tools. The project is archived and no longer maintained.

Agent Chat UI gives developers a web interface for talking to their LangGraph agents and returning to existing conversations. It connects to an agent server you choose, so you can use it with a local development server or a remote deployment.

Wyoming Satellite connects a microphone and audio playback device to Home Assistant for a self-hosted voice assistant. The project is archived and no longer maintained; Linux Voice Assistant replaces it using the ESPHome protocol.

Text Generation Inference (TGI) serves LLMs on your own hardware for developers building chat apps and text generation services. The project is archived and no longer receives updates.

Bytebot carries out tasks described in plain language on its own virtual desktop. It's for people automating document work, research, or tasks across several applications. The project is archived and no longer maintained.

Acontext gives AI agents reusable memory from past tasks. It's for developers who want agents to retain successful approaches, mistakes, and user preferences in files they can inspect and correct.

Verba is a self-hosted document chatbot for people who want to ask questions across their files and knowledge bases. The project is archived and no longer maintained, so it won't receive bug fixes or security patches.

LlamaIndex Chat is a chatbot builder for people who want assistants that answer questions about their own documents. It's an example application built with LlamaIndexTS; you can host the app yourself, but it relies on cloud services for document indexing and AI responses.

MosaicML Composer is a Python library for training neural networks on your own GPUs or multi-node clusters. Built on PyTorch, it's aimed at researchers and engineers who want distributed training with room to customize their experiments. It's open source under Apache 2.0, and prebuilt Docker images are available.

Browser Control lets trusted coding agents work in the Chromium-family browser you already use. It's for developers who want browser automation to use their existing accounts and extensions.

Codestral and Devstral are coding models for developers who want to run completion tools or software engineering agents on their own hardware. The associated mistral-inference library is archived and no longer maintained.

Flowise is a visual builder for developers and teams creating AI agents and chatbots from connected workflow blocks. The project is archived and no longer maintained.

pgai is a Python library for developers building RAG applications and AI agents around PostgreSQL data. The project is archived and no longer maintained or supported.

Void is a desktop AI code editor for developers who want to choose the models behind their coding assistant. It builds on VS Code as a fork; the project is archived and no longer maintained.

TensorZero is a self-hosted platform for developers who need to compare and improve models in LLM applications. The project is archived and no longer maintained.

ollama-telegram is a Telegram bot for people who want to chat with an Ollama model through their messaging app. The project is archived and no longer maintained.