How much AI fits
on your PC?
AI PC Index benchmarks your GPU, shows the local AI models that will actually run on it, and gives you a code to the AI PC Library. One file, one run, no account.
Download AIPCIndex.exeversion 1.0.5 · 61 KB · Windows Have a code? Open the library
bacb524a1b7b766a5c3934b07784acb0a7ddf9f12ce328bcf8c433c52a105f5e
Your GPU's local AI report
Not another synthetic benchmark. A practical answer to “which models can I run?” — with a score you can compare and a code that unlocks the library.
One score
A single number for local AI readiness: GPU class, VRAM and a short compute benchmark.
What will run
LLM, image, speech and video models with minimum VRAM and quantization hints. No guesswork.
Your passport
An anonymous code, one per machine. It opens the library now and your panel later.
A code, not an account
No email, no password, no “sign in with…”. The app issues a short code like XXXX-XXXX after the benchmark. It proves the run happened, opens the AI PC Library, and later becomes the login to your personal panel.
The code is shown even if your GPU is below the full-list threshold. In that case the library simply shows lighter models and CPU options. One code per machine; close the app and the code stays valid.
The code is anonymous: it is tied to a machine ID and GPU, not to your name, email, files or browser.
Code → what it unlocks
From “can I run it?” to “here is the link”
The library is the payoff. Enter your code and get a model list matched to your GPU, with the right quantization and direct links to Ollama, Hugging Face and LM Studio.
Your code
Example: RTX 3060 · 12 GB
Where this is going
Now — AI PC Index 1.0.5
GPU/VRAM detection, AI PC Index score, local model fit, access code, library beta. One file; close it after the code.
Q4 2026 — code login + library v1
The code becomes a login. Full model list, quantization recommendations, direct links, saved machines.
Q1 2027 — personal AI PC panel
Your dashboard: score history, model compatibility, new-model alerts and a “what changed” view for your GPU.
Q2 2027 — AI PC Index Report
An anonymized map of the local AI GPU base for model labs, OEMs and AI tool makers: which models have an addressable audience, where, and at what VRAM.
How many GPUs are actually out there?
Every benchmark is an anonymous data point. The AI PC Index turns those points into a live picture of local AI hardware: GPU classes, VRAM distribution and which models the installed base can run.
For labs: the AI PC Index Report shows addressable GPU base by model, VRAM tier and region. Anonymized, aggregate, no personal data. hello@aipc.digital
What we read, what we don't
We read
GPU model, VRAM, a short local benchmark score, a machine ID and your code. That is what the index and the library need.
We don't read
Documents, photos, browser history, screen, clipboard, keystrokes. The benchmark runs locally; the app is a single file.
You control
Close the app after the code. The code stays valid. The library is a website; no installer, no account, no lock-in.
One file. One minute.
Download for Windows Open the library
AIPCIndex.exe · 1.0.5 · 61 KB · SHA-256 bacb524a1b7b766a5c3934b07784acb0a7ddf9f12ce328bcf8c433c52a105f5e · Windows 10/11.