# Aspen

> Aspen is private AI that runs on your own hardware. Free desktop app for Mac and Windows, free iPhone app, or the optional dedicated Aspen device. No subscriptions, no cloud, no one reading your conversations.

## What is Aspen?

Aspen lets anyone run powerful AI models locally and privately, with no data sent to any company's servers. Download, open, and ask. Core chat and coding work fully offline.

## Key facts

- The desktop app (Mac, Windows) and the iPhone app ("Aspen Local AI") are free, with no account.
- Everything runs on your own hardware — prompts, files, and data never leave your device and are never used for training.
- Runs the latest open models including Llama, Qwen, DeepSeek, Mistral, and Gemma; auto-detects your hardware and recommends a fit.
- Works offline for core chat and coding once a model is downloaded.
- Exposes an OpenAI-compatible API at http://localhost:4000/v1 (plus an optional private tunnel URL); works with the OpenAI and Claude SDKs, LangChain, Cursor, and Continue.dev.
- The Aspen device is an optional, always-on machine for the largest models — about 1 petaflop, 128GB unified memory, up to ~200B parameters. You never need it.

## Documentation

Full documentation: https://runonaspen.com/docs
Machine-readable full text: https://runonaspen.com/llms-full.txt

- Overview: What Aspen is and how it runs AI privately on your own machine. (https://runonaspen.com/docs#overview)
- Install: Get Aspen running on Mac, Windows, or iPhone in a couple of minutes. (https://runonaspen.com/docs#install)
- Quickstart: Open Aspen, pick a model, and ask your first question. (https://runonaspen.com/docs#quickstart)
- Choosing a model: How to pick the right local model for your machine. (https://runonaspen.com/docs#models)
- Chat & artifacts: Conversational AI with live code and HTML previews. (https://runonaspen.com/docs#chat)
- Voice: Hands-free voice conversations, processed on-device. (https://runonaspen.com/docs#voice)
- Vision: Analyze images and screenshots with a local vision model. (https://runonaspen.com/docs#vision)
- Tools: Web search, URL fetch, shell commands, and more — all run on your machine. (https://runonaspen.com/docs#tools)
- Memory (World Model): A private, local memory that personalizes answers. (https://runonaspen.com/docs#memory)
- Privacy: Why Aspen is private by design. (https://runonaspen.com/docs#privacy)
- Developer API: Drop-in, OpenAI-compatible API for your own apps. (https://runonaspen.com/docs#api)
- Connectors (MCP): Connect tools like GitHub through open MCP connectors. (https://runonaspen.com/docs#connectors)
- The Aspen device: Optional always-on hardware for the largest models. (https://runonaspen.com/docs#device)
- Troubleshooting: Common issues and quick fixes. (https://runonaspen.com/docs#troubleshooting)

## Links

- Home: https://runonaspen.com
- Try in browser: https://runonaspen.com/app
- iPhone app: https://apps.apple.com/app/id6775307566
- Source: https://github.com/spideysense/OpenLLM