Asiden runs a language model, voice generation and screen vision on your own PC. That is what keeps your conversations on your machine, and it is why the hardware matters more than it would for a browser-based companion.
These are the same requirements published on the Steam store page.
1. Minimum and recommended
| Minimum | Recommended | |
|---|---|---|
| Operating system | Windows 11, 64-bit | Windows 11, 64-bit |
| Processor | Quad-core CPU | 8-core CPU |
| Memory | 16 GB RAM | 32 GB RAM |
| Graphics | NVIDIA GPU with 8 GB VRAM | NVIDIA GPU with 16 GB VRAM |
| Storage | 25 GB available | 40 GB available |
| Sound | Any Windows-compatible output. Microphone optional, for voice input. | As minimum |
A 64-bit processor and operating system are required. Asiden is built exclusively for Windows 11 — there is no Windows 10 build, and no macOS, Linux, iOS or Android version.
2. Why VRAM is the number that matters
Local models come in different sizes, and the size your machine can handle is decided mostly by VRAM — the memory on your graphics card. It is the same principle as texture quality in a game: a bigger card lets you turn the settings up.
More VRAM means a larger, more capable model, with faster and more detailed replies and more responsive screen vision. The app measures your GPU on first run and recommends a setup that fits it, so you do not have to work this out yourself. You can try a larger model and step back down if your machine finds it heavy — and none of that choice affects privacy. Every local model, large or small, keeps the conversation on your PC.
3. First run needs a connection
An internet connection is required the first time you run Asiden. The app measures your GPU and recommends one Gemma model for both local conversation and screen vision; downloading it takes roughly 4–19 GB, depending on your choice. That is why recommended storage is higher than the installer itself.
After that, everything runs on your machine and works offline. Local chat, screen vision, speech, memory and safety all keep working with no connection, including with Steam in offline mode.
4. Planning your disk space
Allow 25 GB minimum and 40 GB if you want room to keep more than one model. The app and your companion’s memories are small; the models are not. An SSD makes startup and model loading noticeably quicker.
5. If your GPU falls short
The 8 GB VRAM minimum is a real floor for running models locally. If your card is below it, a Plus or Premium plan runs chat and vision through premium cloud models instead of your graphics card, so a thin laptop gets the same conversational quality as a gaming tower.
Worth being clear about the trade: cloud requests are processed off your machine, which is the one thing local mode avoids. The app tells you when a request is going to the cloud, and you can switch back to local at any time. The detail is in the Privacy Notice.
6. Common questions
Will Asiden run on Windows 10?
No. Windows 11, 64-bit only.
Is there a Mac or Linux version?
No, and none is planned at present.
Does an AMD or Intel graphics card work?
The published requirement is an NVIDIA GPU. The app measures your hardware on first run and tells you what it can offer you.
Do I need a microphone?
No. Voice input is optional — you can type normally. Your companion still speaks using local voice generation.
Will it slow down my games?
A demanding game and a large local model compete for the same card. Choosing a smaller model or optional cloud chat can reduce GPU competition, but games, rendering and other local components still use resources. Performance depends on your PC and workload.
Does it work offline?
Yes, after online activation and local setup. Downloads, updates, web searches and optional cloud features need a connection. See what works offline.
Ten dollars, once.
Asiden is a one-off $10 during early access, and local conversations are unlimited after that. Your key arrives by email; enter it in the app after installing.