Local & Self-Hosted AI Tools

Video2X runs AI video upscaling and frame interpolation on your own hardware. It's for people who want to increase a video's resolution or generate extra frames for smoother motion, including those working with anime footage. The project is open source under AGPL-3.0.

Kiln is a desktop workbench for teams building AI applications on macOS, Windows and Linux. It keeps a task and its dataset together across evaluation, prompt optimization, RAG and fine-tuning, so teams can compare changes against the same examples. Engineers, data scientists, QA staff and subject matter experts can contribute through the app.

pgvector adds vector storage and similarity search to Postgres, so developers can keep embeddings alongside application records in a self-hosted database. It suits applications that need to find similar items while retaining SQL queries, joins and transactional guarantees. It runs on Linux, macOS and Windows, with Docker also supported.

Podcastfy turns documents, websites and images into AI-generated audio conversations, with the option to write transcripts using a local LLM. It's an open-source Python alternative to NotebookLM's podcast feature for creators, educators and researchers who want control over the conversation format or need podcast generation inside their own software. It uses the Apache License 2.0.

FastEmbed is a Python library that generates embeddings on your own hardware for semantic search and retrieval-augmented generation (RAG). It's for developers who need to turn text into searchable vectors without relying on a cloud embedding API. It can run on a CPU or use GPU acceleration, and its Apache 2.0 license makes it open source.

Jupyter AI connects AI agents to notebooks through a chat interface inside JupyterLab. It's for people who work in computational notebooks and want an agent to help write code, debug cells, or run notebook work without moving the conversation to a separate app. The extension runs within JupyterLab; the choice of agent determines the AI service it connects to.

Subtitle Edit is an MIT-licensed subtitle editor for Windows, macOS and Linux. It's for people creating captions, translating dialogue or fixing subtitles that don't match the video. Its core editing, conversion and video playback work offline on your device, with optional AI tools for transcription and translation.

MLflow brings agent tracing, LLM evaluation, and model experiment tracking into a platform you can run locally or on your own servers. It's for developers and teams who need to understand failures, compare changes, and monitor AI applications in production. It's open source under Apache 2.0.

HunyuanVideo is an AI video generation model for creators and developers who want to generate footage on their own hardware. Tencent provides model weights and inference code for text-to-video and image-to-video generation, alongside a hosted web experience. Local inference runs on your GPUs; the web offering runs through Tencent's service.

LLM Compressor is an open-source Python library for developers preparing models to run on their own hardware with vLLM. It reduces model size and memory requirements through quantization, and accepts local checkpoints or models from Hugging Face repositories. It's licensed under Apache 2.0.

GFPGAN is an open-source AI face restoration tool for people repairing poor-quality photos and developers adding restoration to image workflows. It runs locally with Python and PyTorch, with Linux support and optional NVIDIA GPU acceleration through CUDA. Its face models use knowledge learned by a pretrained generative model such as StyleGAN2 to reconstruct facial detail.

Surya is a local OCR toolkit for developers extracting text and structure from PDFs and document images. It combines text recognition, layout analysis and table recognition in one vision-language model, so results retain page structure and reading order rather than just the words.

PaddleOCR is an open source OCR and document parsing toolkit for developers building document search, RAG systems and AI agents. It runs on your own hardware or a self-hosted server and turns PDFs and images into structured Markdown or JSON. The Python toolkit uses PaddlePaddle and carries the Apache 2.0 license.

Linkwarden is a self-hosted bookmark manager for people and teams who want to keep the pages behind their saved links. It archives webpages as HTML, screenshots, and PDFs, so you can revisit the content even if the original site removes it. It's useful for research, shared reference collections, and articles you want to read later.

Viseron is a self-hosted network video recorder (NVR) that combines camera recording with local AI computer vision. It's for people monitoring a home, office or other property who want the video system and its analysis to run on their own hardware. Video processing stays local.

MMAudio generates audio that matches a video's action and timing, with text prompts available to guide the result. It also creates audio from text alone. It's for video creators who need sound for silent footage and researchers working on audio generation.

Accelerate is a Python library for developers and researchers who write their own PyTorch training loops and want to use the same code on a local machine or a distributed cluster. It handles the hardware-specific work while leaving the training logic under your control.

Ultralytics YOLO is an open-source Python computer vision library for developers building applications that analyze images and video on their own hardware. It supports local and edge deployment, including NVIDIA Jetson, Raspberry Pi and mobile phones. A separate hosted platform provides browser-based annotation, cloud GPU training and managed prediction endpoints.

ACE-Step is an open-source music generation model for musicians, producers and developers who want to create and edit music on their own hardware. It generates songs with vocals or instrumental tracks from text descriptions and supplied lyrics. You can choose the duration and describe the sound with genre tags, longer prompts or a scene description.

Vespa is a self-hosted AI search platform for developers building search, RAG, and recommendation systems over large, changing datasets. It combines retrieval with machine-learned ranking, so an application can find candidate results and evaluate their relevance in the same platform. The code is open source under Apache 2.0. You can run it on your own servers or use the managed Vespa Cloud service, where applications run in the cloud.

AppFlowy is a self-hosted Notion alternative for teams that want project management, shared documents, and AI in an environment they control. Its self-hosted enterprise offering can run on premises, in your own cloud, or in an air-gapped environment. Self-hosted LLMs and embedding models let you keep AI processing and workspace data inside your infrastructure, with no vendor access to your instance.

KServe is an open source platform for teams serving LLMs and predictive machine learning models on their own Kubernetes infrastructure. It puts both kinds of workloads under a common serving API, so teams can manage different model frameworks through the same platform. It uses the Apache 2.0 license.

WebLLM runs language models directly in a user's browser, using WebGPU for GPU acceleration. It's an open-source engine for developers building web-based AI assistants and Chrome extensions that process prompts on the user's device rather than an inference server. The project uses the Apache 2.0 license.

LobeHub is an AI agent workspace for people who want to assign work to several assistants and keep that work organized. It offers a hosted service and a Docker-based self-hosted version for a private device or server. A Linux download is also available.

LTX-Video is an AI video generation model for creators building controlled animations and developers adding video tools to their own products. You can run it locally or on your own servers using publicly available weights. The LTX family also offers a managed cloud API; local deployments can run in isolated environments without a cloud dependency.

nvtop is a terminal monitor that shows activity across multiple GPUs and accelerators in an interface familiar to htop users. It's useful for people running local LLM workloads or managing compute servers who want to see which processes are using their hardware and how much GPU memory they consume. It runs locally on Linux and is open source under GPLv3 or later.

Haystack is a Python framework for developers building self-hosted AI agents, document search, and apps that answer questions using their own data. Its modular pipelines let teams control which information reaches a model and inspect how retrieval, memory, tools, and generation contribute to an answer. It's open source under Apache 2.0.

NeMo Guardrails is an open-source Python toolkit for developers who need control over how an AI assistant responds and uses tools. It runs within your application or as a self-hosted server, including in Docker. The library uses the Apache 2.0 license.

Sentence Transformers is an open-source Python library for developers building semantic search and document retrieval on their own hardware. It runs embedding and reranker models locally, turning content into numerical representations for similarity comparisons and scoring results against a query. The library uses the Apache 2.0 license.

Ollama Python connects Python applications to models running through Ollama on your own machine or to Ollama's cloud service. It's for developers adding local LLM features to scripts, chat applications, or other Python projects. The library is open source under the MIT license and requires a running Ollama service for local use.

Candle is a Rust machine learning framework for developers who want to embed local AI in applications or deploy models on their own servers. It produces lightweight binaries that don't need Python in production, making it a candidate for serverless inference where a large runtime can slow startup. Its API uses tensor operations familiar to PyTorch developers.

Ultimate Vocal Remover is a desktop app for separating vocals and other stems from audio files on your own computer. It's for people making karaoke backing tracks, isolating a vocal, or working with separate parts of a song. It runs on Windows, macOS and Linux, and its MIT license allows you to use and modify the software.

Steel is a browser API for developers building AI agents that interact with websites. It runs Chrome sessions locally or on a self-hosted server through Docker, so you can keep browser infrastructure on hardware you control. The code uses the Apache 2.0 license. Steel also offers a hosted service where browser sessions run in its cloud.

OCRmyPDF turns scanned PDFs into documents you can search and copy text from while preserving the resolution of their original images. It's a local command-line tool for people digitizing paper records and developers building document processing systems. Your documents stay on your machine.

mergekit combines existing language models into a single model on your own hardware, without additional training or access to the original training data. It's for developers and researchers who want to combine fine-tuned capabilities or adjust the balance between model behaviors. The Python toolkit is open source under LGPL-3.0.

Infinity Embeddings is a self-hosted server for developers building semantic search and retrieval-augmented generation applications. It runs embedding and reranking models on your own hardware, with support for image and audio search alongside text. It's open source under MIT.

Khoj is an AI assistant for people who want to ask questions across their own files, research the web, and give recurring work to agents. You can self-host it on your computer or server, or use Khoj's cloud app. It's open source under the GNU AGPL v3.0 license.

Umi-OCR turns screenshots, images, and scanned documents into text on Windows and Linux. It works offline. The software is free under the MIT License, and recognition runs locally with built-in language libraries. It suits people who need to copy text from a screen, process a folder of images, or make scans searchable without sending them to an online OCR service.

Upscayl enlarges low-resolution photos and graphics with AI. Its free, open source desktop app runs locally on Linux, macOS, and Windows, making it an option for people who want to improve images on their own computer. It needs a Vulkan-compatible GPU and is licensed under AGPL-3.0.

Vane (formerly Perplexica) is a self-hosted AI search engine for people who want answers drawn from web results, with citations they can check. It runs on your own hardware through Docker or as a server application, with a browser interface and locally stored search history. The project is free and open source under the MIT license.

Msty Studio is a private AI workspace for chatting with local or hosted models and working with your own documents. It offers a desktop app alongside web and team deployment options. Local model support includes MLX, llama.cpp and Ollama.

OpenHands is a coding agent platform for developers and teams that want agents to make changes across a codebase and handle recurring engineering work. You can run it on your own computer or self-host it on a server. Its core is open source under the MIT license, and a separate OpenHands Cloud service offers hosted execution.

Docker Model Runner lets developers run and serve AI models on their own computer or server using Docker Desktop, Docker Engine or the standalone dmr binary. It pulls models from Docker Hub, OCI registries, and Hugging Face, then stores them locally. Inference runs locally too.

faster-whisper is a Python library for people building local speech transcription into their own software. It runs OpenAI's Whisper models through CTranslate2, with faster processing and lower memory use than the original Whisper implementation in the project's comparisons. It runs on a CPU. NVIDIA GPUs are supported too, and the code is open source under the MIT license.

Vercel AI SDK is a TypeScript library for developers building chatbots, generative interfaces, and AI agents. It gives an application one way to work with models from providers such as OpenAI, Anthropic, and Google, so teams can change providers without rebuilding the parts of their app that handle responses. It suits developers who want control over the application they build while choosing how it connects to model services.

FLUX.1 is a family of image models for people who want to generate or edit images on their own infrastructure. Black Forest Labs provides Python inference code for its open-weight models and a separate hosted API. Local inference runs on your hardware; API requests go to Black Forest Labs.

Gradio turns a Python function, model, or API into a web interface. It runs on your computer, making it useful for developers and researchers who want people to try an AI project through a browser. It's open source under the Apache 2.0 license.

Krita AI Diffusion brings image generation into Krita for artists who want to paint and edit in the same workspace. The open source plugin is licensed under GNU GPL v3.0 and can run generation for free on your own hardware through ComfyUI. Interstice.cloud offers a separate online generation service for people without a suitable GPU.