Repolode
Submit
off-grid-ai

off-grid-ai/OGAD

A private desktop AI studio for macOS, Windows and Linux that runs open models locally, with an OpenAI-compatible gateway and no account or cloud.

Open-source alternative to:LM StudioOllama

View on GitHubWebsite
Stars
131
Forks
25
Open issues
15
Last commit
30 minutes ago

What is it?

Off Grid AI is a desktop application that runs open models entirely on your own machine. It covers chat with text and vision, on-device image generation, speech-to-text and text-to-speech, and a canvas that renders HTML, React, SVG and Mermaid in a sandbox. Inference is handled by bundled llama.cpp, stable-diffusion.cpp, whisper.cpp and Kokoro, so conversations, files and models stay on the device.

The app also exposes a local OpenAI-compatible gateway at 127.0.0.1:7878 for chat, vision, images, audio and embeddings. It can run headless, with model pull, activate and delete managed over HTTP. Projects support RAG over uploaded documents with cited sources, and MCP connectors can be added for in-chat tools.

An optional Pro license adds an always-on layer that captures and OCRs the screen, builds a journal and entity records, transcribes meetings, and provides dictation and a clipboard history, with actions gated by user approval. The open core is AGPL-3.0 licensed; Pro lives in a separate private package.

Key features

  • 01Chat with text and vision models, including a reasoning mode.
  • 02On-device text-to-image and image-to-image using SDXL and Z-Image-Turbo models.
  • 03Speech-to-text and text-to-speech with a hands-free voice mode.
  • 04OpenAI-compatible gateway for chat, images, audio and embeddings.
  • 05Projects with RAG over PDF, DOCX, txt and md files, citing sources.
  • 06MCP connector support for tools used inside chat.

Use cases

  • Chatting with and analyzing documents privately on a laptop without sending data to a cloud service.
  • Running a local model on a homelab server as an OpenAI-compatible endpoint for your own apps.
  • Generating images or transcribing audio offline on a Mac with Apple Silicon.
  • Prototyping LLM-backed tools against a local API that mirrors the OpenAI interface.

Quick start

  1. 1

    Download the app from the releases page

  2. 2

    Clone the repository

    git clone https://github.com/off-grid-ai/OGAD.git
  3. 3

    Enter the project folder and pull the bundled native binaries with Git LFS

    cd OGAD && git lfs install && git lfs pull
  4. 4

    Install dependencies and start the full app

    npm install && npm run dev

Things to know

  • Desktop builds cover macOS (Apple Silicon), Windows (x64) and Linux (x64, Ubuntu 24.04 or newer) only.
  • Building from source requires Git LFS to pull the bundled native binaries.
  • Off Grid AI Pro is a separate paid license and is not part of the open-source core.

Topics

#local-ai#llm#on-device-ai#openai-compatible#rag#mcp#privacy#offline-ai#image-generation#ai-assistant

FAQ

What is OGAD?
A private desktop AI studio for macOS, Windows and Linux that runs open models locally, with an OpenAI-compatible gateway and no account or cloud. Off Grid AI is a desktop application that runs open models entirely on your own machine. It covers chat with text and vision, on-device image generation, speech-to-text and text-to-speech, and a canvas that renders HTML, React, SVG and Mermaid in a sandbox. Inference is handled by bundled llama.cpp, stable-diffusion.cpp, whisper.cpp and Kokoro, so conversations, files and models stay on the device.
Is OGAD free to use? What license does it have?
OGAD is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
Is OGAD still maintained?
Actively maintained. The last commit to OGAD was on Oct 22, 2026.
What can I use OGAD for?
Chatting with and analyzing documents privately on a laptop without sending data to a cloud service. Running a local model on a homelab server as an OpenAI-compatible endpoint for your own apps. Generating images or transcribing audio offline on a Mac with Apple Silicon. Prototyping LLM-backed tools against a local API that mirrors the OpenAI interface.
How do I install OGAD?
Download the app from the releases page Clone the repository: git clone https://github.com/off-grid-ai/OGAD.git Enter the project folder and pull the bundled native binaries with Git LFS: cd OGAD && git lfs install && git lfs pull Install dependencies and start the full app: npm install && npm run dev
Where can I get OGAD?
The source code and installation instructions are on GitHub: https://github.com/off-grid-ai/OGAD

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects