Chat with an AI model completely offline once the initial model download finishes.
Ask questions about a folder of your own PDFs or notes without uploading them anywhere.
Write stories, roleplay with characters, or build branching dialogue trees.
Generate finished PDFs, code files, or Markdown documents saved directly to your computer.
| penpro/aphelion | alamops/agetor | aza-ali/blendpixel.com | |
|---|---|---|---|
| Stars | 12 | 12 | 12 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | easy | hard | easy |
| Complexity | 2/5 | 4/5 | 2/5 |
| Audience | general | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
First launch downloads an AI model sized for your hardware (roughly 2 to 16 GB), a GPU is recommended but not required.
Aphelion is a desktop application that runs an AI language model entirely on your own computer, with no account, cloud service, or internet connection required after the initial setup. It works on Windows, macOS on Apple Silicon, and Linux, and once a model is downloaded the app can keep working with the internet completely disconnected. The name refers to the point in an orbit farthest from the sun, meant to suggest an AI as far as possible from the cloud. Inside the app, people can chat with characters for roleplay, write stories, build branching dialogue trees, or talk to an assistant tuned for tasks like coding help or describing images. It can also read a folder of PDFs or notes you point it at and answer questions based on that material, without those files ever leaving your machine. Beyond chat, it can generate finished documents such as PDFs with tables and formatting, or code, HTML, and Markdown files saved directly into your folders, and it can open and edit existing files too. The README is careful to explain that Aphelion is not an autonomous agent. The AI model only reads text you give it and writes text back. It cannot run programs or commands, browse the internet on its own, control your computer, or touch any file or folder unless you specifically grant access to it. The app itself handles reading files you approve and saving documents you request, but there is no ongoing loop where the AI decides to take actions by itself, and its local server only listens on your own machine. Installing it is meant to be simple: download the installer for your operating system from the releases page, run it, and on first launch the app examines your hardware and downloads an AI model sized to fit it, usually a few gigabytes. A graphics card with more video memory allows a larger, more capable model, though the app can also pick something smaller for weaker hardware. The project is released under the MIT license and macOS and Linux support is newer and less tested than the Windows version.
A free desktop app that runs an AI model entirely offline on your PC for chat, roleplay, story writing, and document generation.
Mainly TypeScript. The stack also includes Tauri, llama.cpp, TypeScript.
Licensed under MIT, which allows free use, modification, and distribution, including commercially, as long as you keep the license notice.
Setup difficulty is rated easy, with roughly 30min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.