Aspen / Guide
Local AI: run AI on your own computer
Local AI means the model that answers you runs on hardware you control instead of a remote cloud service. Once a model is downloaded, core chat and coding can run on-device and offline. Aspen packages the model, inference engine and interface into a normal app so you do not need to manage a terminal or server.
What local AI actually changes
With cloud AI, your prompt travels to a provider's infrastructure, the model runs there, and a response comes back over the internet. With local AI, inference happens on your Mac or PC. That changes the privacy boundary, cost model and availability of the system.
- Privacy: core prompts and files can stay on your own device.
- Offline use: after the model is downloaded, chat and coding do not require a connection.
- No per-token meter: you use your own compute rather than paying for every request.
- Control: you can choose among open models and change them as your needs change.
What local AI is good at
Modern open models are useful for writing, summarizing, brainstorming, coding help, document work and many everyday questions. They are also useful when the input is sensitive or when internet access is unreliable.
The tradeoff is straightforward: the largest cloud models still have an advantage on the hardest reasoning tasks, while a local model is limited by the memory and compute available on your machine. Many people use both: local AI for routine and private work, and a frontier cloud model when they explicitly need maximum capability.
What hardware do you need?
A modern laptop or desktop is enough to start. As a rough guide, 8GB of memory supports small models, 16GB is comfortable for many 7–8B models, 32GB opens up 13–14B-class models, and higher-memory machines can run larger options. Quantization reduces memory needs. Apple Silicon is particularly convenient because CPU and GPU share unified memory.
Aspen checks your hardware and recommends models that fit rather than asking you to reason about parameter counts and quantization files yourself.
How to start with Aspen
- Download Aspen for Mac or Windows, or install Aspen Local AI on iPhone.
- Open the app and choose a recommended model.
- Download the model once.
- Chat, code or work with your local model. Core use continues to work offline.
No account or subscription is required for the app. Optional online tools such as live web search still need a network connection.
Local AI vs local LLM
Local AI is the broader category: an AI capability running on your own hardware. A local LLM is specifically a large language model running locally. Aspen uses local LLMs for chat and coding, and exposes them to other software through an OpenAI-compatible local API.
Frequently asked questions
Does local AI work without internet?
Yes for core inference after the model is downloaded. Internet is still needed to download new models or use explicitly online tools such as live web search.
Is local AI free?
Open-weight models can be downloaded and run without per-token fees. Aspen’s Mac, Windows and iPhone apps are free; your machine supplies the compute and electricity.
Is local AI as good as ChatGPT?
For many everyday writing, summarization and coding tasks, modern local models are strong. The largest cloud models still lead on some difficult reasoning tasks, so the right choice depends on the job.
Run AI on your own hardware
Aspen is free on Mac, Windows and iPhone. No subscription or account required.
Get Aspen free →