Local & Self-Hosted AI Tools

OmniRoute is an AI gateway for developers who want their coding assistants and other AI clients to share one set of provider connections. It routes requests across models and accounts when a service fails, reaches a rate limit or runs out of quota.

Upscayl is an AI image upscaler for people working with small photos, illustrations or other low-resolution images. Its desktop app processes images locally, with a separate cloud offering for work across devices.

Paperclip coordinates AI agents for people managing work across development, research, marketing and other business functions. It gives the team a shared structure and keeps its work tied to organization goals.

Cua supplies desktops and automation tools for developers building AI agents that operate native apps and browsers. You bring your own agent and model; Cua provides the computer environment and ways to interact with it.

Agent Reach gives AI agents access to online content for research, repository work and social media searches. It's for people using Claude Code, OpenClaw, Cursor, Windsurf or another agent that can run shell commands.

LangWatch helps AI teams test agent behavior before release and investigate production failures. Teams can reproduce a failed conversation and check how an agent handles it again.

CC Switch is a desktop manager for developers who use multiple AI coding assistants or API providers. It brings their settings into one interface and preserves existing plugins, hooks and other custom settings when switching providers.

vLLM serves language models on hardware you control. It's for developers and teams building model APIs that need to handle concurrent requests while using memory and compute efficiently.

FalkorDB is a self-hosted graph database for developers building GraphRAG applications and AI agents that retrieve connected facts. It stores entities, relationships and their attributes, with Cypher queries and vector search for finding relevant information. Docker runs the database on your own hardware; a separate hosted cloud service is also available.

Docmost is a self-hosted team wiki with collaborative editing, spaces, page history, attachments and access permissions. Teams can run the wiki on their own servers or use managed hosting. Its core is AGPL-3.0 licensed; AI search and chat require a separate commercial license.

Supergateway connects MCP tool servers to AI clients that expect a different connection type. Developers can expose a local server over a network or make a remote server accessible to a desktop client.

OpenAI Agents SDK is a Python framework for developers building tool-using agents with local models or cloud APIs. It supports gpt-oss through Ollama, with LM Studio also documented as a local model option. OpenAI models use the remote Responses API by default and require an API key, even when the agent runtime and workspace run on your machine.

LightLLM is a self-hosted LLM inference and serving framework for developers and teams running models on their own GPU servers. Its focus is serving requests efficiently: it schedules work around predicted memory use and handles text encoding, model inference and output decoding asynchronously to improve GPU utilization.

Vellium is a desktop AI app for writers and people who want conversations with fictional characters. It combines multi-character roleplay with writing projects and lets you choose the model provider.

AIClient2API is a self-hosted AI API proxy for developers connecting chat apps and coding assistants to multiple model providers. It exposes services such as Antigravity, Codex, Grok, and Kiro through interfaces those apps already support.

text-generation-webui, also called TextGen, runs language models on your own hardware for chat, writing and application integrations. You can use it as a desktop app or a self-hosted browser interface.

Optimum Quanto is an open-source PyTorch library for developers who want to reduce the memory needed to run models on their own hardware. It handles language models and image-generation models, with integrations for Hugging Face Transformers and Diffusers. The project is in maintenance mode, accepting minor fixes and documentation work rather than major features.

My Virtual Office gives people running AI agents a visual way to follow their work. Agents appear as characters in a retro 2D office, so you can see who's active without relying only on logs.

Exo Agent Framework is for developers building long-running AI agents that can revise their own code, tools and prompts. It preserves agent identity and history as their behavior changes across runs.

DeepSqueak is a MATLAB application for bioacoustics researchers studying animal vocalizations, including ultrasonic calls from mice and rats. It uses neural networks to detect calls in audio recordings and provides a graphical interface for reviewing detections. The code is open source under the BSD-3-Clause license.

LocalSheets is a spreadsheet for people who want their workbooks on their own disk, with optional AI through a local Ollama instance. It runs from a single HTML file in a desktop browser and works offline without an account or cloud service. The app doesn't collect analytics or tracking data.

Kiro Crew is a self-hosted AI development workspace for developers who want agents to keep working between conversations. It runs on your computer or a server you control, but its default kiro-cli backend requires separate Kiro sign-in for model access. The default Kiro connection sends model requests to Kiro's cloud service, powered by Amazon Bedrock.

Research Agent is a free, MIT-licensed, open source AI agent that combines online research with documents on your machine. It's for people investigating technical topics or comparing papers who want cited reports and a choice of local or cloud models. It runs on macOS, Windows and Linux, with a browser dashboard and a command-line interface for automation.

localGPT is a self-hosted document chat app for people who want to ask questions about files and summarise them on their own hardware. Its local processing keeps document content and chat data on the machine running the app.

Gemini CLI is a coding assistant for developers who want to work with an AI agent in their terminal. It can inspect a codebase and act on files and shell commands, with Gemini handling the model responses in Google's cloud.

SilverBullet is a knowledge workspace for personal notes, shared wikis and project documentation. It keeps your writing in plain Markdown files and lets you build connections and collect information across pages.

MioSub is an AI subtitle editor for creators and translators working with video or audio. It handles subtitle generation and translation for local files or downloaded media, including podcasts, radio recordings and audiobooks.

Local Agent Studio is a Windows desktop app for building AI agents and connecting them in visual workflows. It suits first-time agent builders and developers prototyping private automation. Local and cloud agents can work together in the same workflow, with progress visible for each step.

FlashAttention is an attention library for developers training or running transformer models on their own GPUs. It computes exact attention while using less memory than standard PyTorch attention, allowing longer sequences.

DeepXDE is an open-source Python library for researchers and engineers who use neural networks to solve scientific problems. It combines physics-informed neural networks (PINNs) with operator learning, so you can work with equations and physical constraints alongside training data. Its license is LGPL-2.1.

xberg, formerly Kreuzberg, is a local document extraction engine for developers building AI search and document processing applications. It turns files into text and structured content that those applications can use.

ExecuTorch is an on-device AI runtime for developers adding PyTorch models to apps and embedded devices. It handles inference on the user's hardware across Android, iOS, macOS, Linux and Windows, as well as embedded systems and microcontrollers.

HomeGallery is a self-hosted photo and video gallery for people browsing a personal media archive without reorganizing their files. It's designed for a single user with some technical knowledge. By default, AI discovery sends preview images to a public analysis API; you can host that API locally instead.

Pallaidium is a Blender add-on for filmmakers and AI artists who want to develop scenes with generated visuals and sound before shooting. It puts generation on the Video Sequence Editor timeline, where you can test dialogue timing and edit scenes alongside music.

Nemotron is NVIDIA's family of AI models for developers building agents that reason, write code and use tools. Teams can evaluate and adapt the models using published weights, training data and training recipes.

Context Mode is a local MCP plugin for developers whose AI coding agents fill their context windows with logs, file reads and search results. It stores large tool outputs on your machine so the agent can retrieve relevant passages as needed.

CTranslate2 is an inference library for developers who want to run Transformer models on their own hardware or servers. It focuses on reducing execution time and memory use for applications built in C++ or Python.

Notely Voice turns speech and recordings into editable notes for lectures, meetings and interviews. It runs Whisper on your device, so speech recognition works offline after you download a model and doesn't upload audio to a cloud service.

Lithe is a desktop IDE for Java and Spring Boot developers who need to inspect, edit, and run code, including work produced by AI coding assistants. It's an IntelliJ IDEA alternative designed to keep baseline memory use low.

NocoBase is a self-hosted platform for teams building business applications for themselves or clients, including CRM, help desks, and project management systems. It combines a visual no-code editor with support for AI coding agents, so business users can adjust what agents build without editing code. Business data stays in your own database.

HarnessX is a Python framework for developers building AI agents whose behavior they can reuse and adapt. It separates model choice from tool use, memory and context handling, so changing an agent's role doesn't require rewriting it.

CogView4 is a text-to-image model for image creators and developers who want to generate images on their own hardware. It accepts Chinese and English descriptions and can include Chinese text in the results.

AutoKeras is a Python AutoML library built on Keras for automated deep learning. It searches for neural-network configurations and trains models from your data, reducing the amount of manual architecture design needed for an experiment.

MTEB helps researchers and developers compare embedding models and retrieval systems for the tasks they need. Its benchmarks cover multiple languages and modalities.

Kubeflow is a self-hosted AI platform for teams developing and running machine learning workloads on Kubernetes. Its modular projects cover interactive development, training and the work of moving models into production.

Demucs separates mixed music recordings into parts for musicians who want to work with individual instruments or vocals. The author maintains a separate repository, with no active feature development and slow responses expected.