Local & Self-Hosted AI Tools

Agent Lightning is a Python framework for developers and researchers training AI agents through reinforcement learning. It keeps the agent's existing tools, context and task environment involved in training, so feedback comes from the workflow the agent actually uses.

Local OpenAI LLM for Home Assistant connects conversation agents and AI tasks to models served through an OpenAI-compatible API. It's for Home Assistant users who want to choose their own model server for voice conversations and generation tasks.

DeepPoseKit is an open-source Python toolkit for researchers measuring animal movement and behavior in images or video. You define the body points you want to track and train a model on your own annotations. It runs through TensorFlow and Keras on your own hardware, with Windows support documented, and uses the Apache 2.0 license.

Interpreter translates Japanese game text into English subtitles over the game window. It's aimed at retro-game players, with MeikiOCR tuned for Japanese game text and pixel fonts. The default MeikiOCR and Sugoi V4 models run locally without internet access after setup.

LocalDocRAG is a self-hosted PDF question-answering app for people who want to search their documents with a local LLM and check answers against specific pages. It runs through Docker on macOS, Windows, Linux or a Raspberry Pi, with a browser interface accessible from phones and computers on the same network. It's open source under the MIT license.

RealChar lets you create AI companions with their own personalities and hold live voice or text conversations. It's for people who want to shape a character and host the app themselves while using external AI services for replies.

arxiv-mcp-server gives AI assistants tools to find and study arXiv papers. It's for researchers who want an assistant to inspect specific sections and work with a locally stored paper library.

SpeechPulse is a dictation and transcription app for people who want to write by voice or turn recordings into text. Its Whisper speech recognition runs locally, so dictation can work without an internet connection.

ACP Bridge lets teams and bots access coding agents running on a development machine or server through a shared HTTP API. It connects agents through the Agent Client Protocol (ACP), so callers don't need each agent installed on their own computer.

Minutes is a meeting notetaker for people who want searchable records of calls, voice memos and everyday dictation on their own computer. You can read the notes in the app or let an AI assistant search the conversations you choose to share.

ResearchNotebookLM is a free, self-hosted alternative to Google NotebookLM for people studying papers, collecting research or working with private documents. It runs through a browser on Windows, macOS or Linux, and can run on your own server. With Ollama, questions and answers stay local and work offline without a cloud account.

LiveAgent is a desktop AI agent for people who want an assistant to work with files and commands on their computer. Local tool execution is separate from model access, which uses Claude, Codex or Gemini protocols, or services compatible with those protocols through a custom API address.

Paddler lets product and infrastructure teams serve AI models on their own servers and spread requests across multiple machines. It's built for applications that need text generation or embeddings while keeping inference within infrastructure the team controls.

Jina Reader is a web extraction API for developers building AI agents, search tools and systems that answer questions from retrieved documents. It turns page content into Markdown or JSON instead of passing raw HTML to an LLM.

SolidGPT is an AI search assistant for developers, available as a VS Code extension or a self-hosted server with a browser interface. The application can run on your machine or server, but its AI processing uses OpenAI's cloud API. It needs an API key and an internet connection, so hosting it yourself doesn't make model inference local.

OpenVoice generates speech in a reference speaker's voice from a short audio clip. It's for developers and creators who need the same voice across languages or want to vary its delivery without changing the speaker.

Claude DevTools is a local debugger for developers who want to understand what happened during a Claude Code session. It reads saved logs from terminal and IDE sessions without modifying Claude Code.

Oh My Pi is a terminal coding assistant for developers who want an AI agent working directly in their codebase with IDE tools. It's a fork of Pi that supports both local model servers and cloud providers.

sonible smart:chain is a proprietary AI mixing plug-in for producers and audio engineers who want help processing individual tracks and building a rough mix. It runs in a DAW on macOS or Windows, with native Apple Silicon support. Its processors share an analysis of the audio, so their suggested settings account for the rest of the chain.

MLC LLM is a compiler and deployment engine for developers running language models on their own hardware or building AI into apps. Its shared MLCEngine supports deployment across desktop computers, mobile devices and browsers.

TranslateLocally translates text on your own computer for people who want local machine translation and teams working with their own translation models. It uses Marian and Bergamot, with a desktop interface for everyday text and command-line access for batch work.

sherpa-onnx is a toolkit for developers building speech features that work offline on their own devices. It covers transcription, speech generation and audio analysis, with processing kept local.

Torrix is a self-hosted LLM monitoring tool for teams investigating AI spending and failed requests. It keeps request records in your infrastructure and brings usage analytics and individual traces into one dashboard.

Lightning Pose is an open-source package for researchers who estimate animal poses from video for behavior analysis. It supports 2D and 3D pose estimation with single-camera and multiple-camera datasets, including footage where body parts become hidden from view.

Megatron-LM is a Python framework for researchers training large language models on NVIDIA GPU infrastructure. It provides reference training scripts for experimenting with models and studying distributed training.

PaddleX is a low-code toolkit for developers building AI applications with PaddlePaddle models. It brings pretrained pipelines and model training into one tool for applications that process documents, images or time-series data.

Evidently helps AI teams check output quality and catch regressions in chatbots, RAG applications, AI agents and predictive models. Its Python library handles text and tabular data for development evaluations and production monitoring.

Pi Desktop is a desktop workspace for developers who want AI coding agents independent of their editor or terminal. It keeps projects and agent sessions together, with tools to review changes and inspect ongoing work.

MiniMax-M1 is a reasoning model for developers building AI agents or working on coding and mathematical problems with long inputs. You can deploy it on your own infrastructure.

PentestGPT is an AI agent for security testers working on authorized penetration tests and capture-the-flag (CTF) challenges. It can carry out a test autonomously or let you guide the work in an interactive mode.

LangExtract is a Python library for developers who need to turn notes, reports or other text into structured data. It links extracted details to their original passages so people can check the results against the document.

M3E Canvas is a visual editor for people planning Material 3 Expressive apps with an AI coding assistant. It lets you sketch screens and test their flow before handing a text prompt to a tool such as Claude Code, Codex, Gemini CLI or Cursor.

Atlas is a desktop workspace for developers who use coding agents across projects and want to keep the context behind their code changes. It brings agent conversations, code review and project history into one window.

Code Review Graph helps developers review changes with AI coding assistants without repeatedly supplying an entire repository. It keeps a local map of code relationships and returns the context relevant to a change or question.

Eidon AI is a self-hosted AI assistant for people who want everyday chat and agents that can carry out longer tasks on their own computer or server. A Chief of Staff agent answers directly, delegates work to specialists, or creates an agent for a task. Agents can message each other and keep working on a schedule.

Minima lets you ask questions about your own documents, with a choice of local inference or an external model. It's a self-hosted RAG tool for people and teams who want document-based answers while controlling where indexing and AI processing happen. It runs in Docker containers and is open source under the Mozilla Public License 2.0.

OneRAG is a self-hosted RAG framework for developers building chatbots that answer questions using their own documents. It combines document retrieval and AI answers, with replaceable components for different models and search needs.

FUTO Voice Input adds private dictation to compatible Android keyboards and apps. It's for people who want speech recognition on their phone while keeping the keyboard they already use. Gboard and Samsung Keyboard don't support it as an alternative voice input provider.

n8n-MCP connects AI assistants to n8n so automation builders can create and maintain workflows through tools such as Claude, ChatGPT and Cursor. It's a connector for an assistant you already use, rather than a standalone AI agent.

llm-d is a distributed LLM inference stack for infrastructure teams serving models across a Kubernetes cluster. It coordinates model servers such as vLLM and SGLang to handle production traffic across multiple machines.

Better Harness helps developers and teams assess how coding agents carry out work. It examines the workflow around a code change to find unclear goals, missing validation and delivery checks that a diff review can miss.

Accentize SpectralBalance is a proprietary dialogue EQ plugin that runs on macOS and Windows, with native Apple Silicon support. It's for dialogue editors, ADR mixers and podcast producers who need speech recorded across different takes or environments to sound consistent.

MLForecast is an open-source Python library for data scientists and developers who need to forecast many time series with machine learning. It combines time series feature engineering with scikit-learn-compatible models, so teams can use regressors such as LightGBM and LinearRegression rather than adopt a dedicated forecasting model. Its license is Apache 2.0.

Ente Photos is a Google Photos alternative for people who want to back up and share their library without letting the storage provider read its contents. You can self-host its server or use Ente's hosted service.

kvpress is a Python library for developers and researchers reducing the memory needed for long-context LLM inference. It compresses the key-value cache that grows as a model processes longer inputs.

Stash gives AI tools persistent memory so preferences, plans and working context can carry into later conversations. It's for people switching between AI tools and developers building agents that need information to survive between sessions.