Self-Hosted AI Tools You Can Run in Docker

PasteGuard is a privacy proxy for teams that use AI with customer records, client files or project secrets. It runs locally or on your own server and checks requests before they reach a model provider.

OpenShell is a self-hosted runtime for teams that need autonomous AI agents to work with files and APIs under explicit access limits. It runs agents in isolated sandboxes and enforces policies at the kernel level.

PageLM is a self-hosted AI study platform inspired by NotebookLM. Students, educators and researchers can use it to turn their materials into resources for revision and exam practice.

FluxGym is a browser interface for training FLUX LoRAs on your own images. It's for people who want to customize image models locally, with support for GPUs with 12GB, 16GB or 20GB of VRAM.

Nomic Embed Text v1.5 is an English text embedding model for developers building search and RAG applications. It represents text numerically so applications can find passages with similar meanings.

RealtimeSTT is a Python library for developers building voice assistants, dictation tools and apps that turn live audio into text. It can show text while someone speaks and produce a final transcript when they finish.

GPT Pilot is a local AI coding agent for developers who want to build applications under human supervision. The project is no longer maintained, and its model calls rely on external providers.

mcp-proxy connects AI clients and Model Context Protocol (MCP) servers that use different connection types. It's for developers who need desktop clients to reach remote tools, or want other clients to access a local MCP server.

All-in-One Sandbox gives AI agents a place to browse websites, execute code and work with files. It's for developers building agents that need these tools together, with a shared filesystem instead of separate environments for each task.

mcpo is a self-hosted proxy for developers connecting MCP tools to AI agents and apps that accept OpenAPI servers. It exposes existing MCP servers through HTTP, avoiding a separate adapter for each tool.

Haystack is a Python framework for developers building AI agents, document search and applications that answer questions using their own data. Its pipelines give teams control over the information a model receives and the steps that produce an answer.

IntentKit is a self-hosted AI agent framework for developers who need several agents to share work and carry out recurring tasks. It runs as a server application, with chat channels and APIs for interacting with the agents.

TTS WebUI is a self-hosted browser interface for people creating speech and music or processing audio with AI models. It brings these tasks into one place, with model choices for different kinds of audio work.

Sandcastle is a TypeScript library for developers building automated coding workflows in scripts, CI pipelines, or custom tools. It coordinates sandboxed AI coding agents and manages where their Git changes land.

Facet combines photo culling with a self-hosted gallery for photographers choosing keepers from a shoot or organizing a personal library. Its AI analysis runs on your own machine without cloud services, accounts or API keys.

llama-swap is a local AI proxy for people who want several models available to their apps without keeping them all loaded. It starts the server a request needs and replaces the active server when another model is requested.

PAL MCP Server connects coding assistants to other AI models for second opinions, code review and planning. It's for developers who want to consult several models within the same coding session, using clients such as Claude Code, Codex CLI, Gemini CLI or Cursor.

Agno is a framework and runtime for developers building customer-facing or internal AI agents. It combines agent development, service hosting and a web interface for managing runs on infrastructure you control.

Presidio is a self-hosted framework for developers and organizations that need to remove personally identifiable information (PII) from data. It detects sensitive details such as names, credit card numbers and phone numbers, then applies configurable de-identification.

Hivemind is a Python library for researchers and teams that want to train a shared neural network across computers owned by universities, companies or volunteers. It uses PyTorch and connects participants over the internet, so contributors can pool their machines for a training job without a central master node.

Vald is a distributed vector search engine for developers building similarity search into applications. It runs on a self-hosted Kubernetes cluster and finds approximate nearest neighbors in dense vectors using NGT. It's open source under the Apache 2.0 license, with an architecture designed to spread large vector collections across multiple machines.

Code Buddy is a terminal AI coding agent for developers who want it to read a repository, make changes and run commands on their machine. It supports local models through Ollama as well as cloud providers.

Linkwarden is a bookmark manager for people and teams building a lasting collection of web references. It saves copies of pages so you can revisit their content after the original links disappear.

CamoFox Browser is a self-hosted browser server for developers whose AI agents need to interact with sites that detect browser automation. It builds on Camoufox, a Firefox fork with browser fingerprint spoofing, to reduce detection by systems such as Cloudflare and Google's bot checks.

RuoYi AI is a self-hosted AI assistant platform for teams building assistants that answer questions about their documents or perform tasks with external tools. It has separate user and admin interfaces.

Wav2Lip changes a face's mouth movements in an existing video to match supplied speech. Its research version is for noncommercial projects, including experiments with lip-sync generation and model training.

MCPJam helps MCP server developers check how their tools and apps behave across AI clients before deployment. It combines hands-on debugging with repeatable tests for differences between ChatGPT, Claude, Cursor, and Copilot.

Sonar serves language and multimodal models for developers and teams running AI applications on their own hardware. It builds on vLLM and works with Hugging Face-compatible models.

ToolJet is an internal-app builder for developers and operations teams creating dashboards, admin panels and business tools around existing data. It can run on your own infrastructure with Docker or Kubernetes, or through ToolJet Cloud. The community foundation uses AGPL-3.0; Enterprise features use separate commercial terms.

DevOpsGPT is a self-hosted AI development application for teams turning written requirements into code and deployment work. The browser interface and generated-code workspace run on your machine or server, while GPT requests use the OpenAI API with a token and incur provider charges.

sandboxd is an AI app builder for teams hosting coding agents, playgrounds or app previews on their own server. It turns prompts into running apps, each with an isolated workspace and its own preview URL.

cook coordinates coding agents for developers who want repeated reviews or competing implementations of a task. It supports Claude Code, Codex and OpenCode, and lets you combine sequential work with parallel attempts.

Taskuary is a personal AI work manager for people whose requests arrive across email, chat and ticket systems. It gathers those requests into tasks and lets AI agents prepare the work for your review.

TensorRT-LLM is an inference library for developers serving LLMs on their own NVIDIA GPUs. It supports application-level inference and server deployments, with performance optimizations for workloads spread across one or several machines.