I kept answering the same Discord and email question: can my GPU run a local LLM? The honest answer always needed GPU VRAM, system RAM, and which model family you meant. So I stopped writing one-off replies and shipped a free local LLM hardware checker on mustafa.net with Cursor — pick a GPU and RAM, get a fit list, then jump into the deeper guides.
Live tool: Can your PC run a local LLM?
Why a checker instead of another matrix post
I already publish VRAM matrices and GPU budget guides. People still bounced because a table is not a decision. They want to select their card, see which models fit, and move on. The checker is that front door. Cursor handled the interactive UI and the wiring into the existing Local LLMs hub; I kept the model/hardware judgments grounded in the same companion pages I already maintain.
What it does today:
- GPU + system RAM selectors (including CPU-only)
- Model-type filters (chat, coding, vision, flagship)
- Fit results with community-style speed estimates — not lab-certified benchmarks
- Links into the deeper posts: VRAM requirements, GPU compatibility, best GPU by budget, quantization
I am not inventing new throughput numbers in this write-up. The tool labels estimates as community-reported and notes that quantization, context length, and backend (Ollama, llama.cpp, vLLM) change real speed.
How Cursor helped me ship it
This was a Cursor one-shot style build on top of the site, not a greenfield SaaS:
- Spec the inputs people actually know (GPU name / VRAM bucket + RAM).
- Reuse the model catalog logic already sitting behind the Local LLMs guides.
- Keep it browser-side — no signup, no API key, no account.
- Soft-link the next step: if the answer is “buy a better card,” send them to Homelab Gear instead of dumping affiliate spam in the checker itself.
Cursor was the pair-programmer for the UI and copy loops. The product rule stayed mine: honest fit answers first, shopping second.
What I left out on purpose
- No login wall
- No “upload your dxdiag” flow
- No fake precision on tokens/sec
- No Laya / agent runtime story — this post is only about the checker
If your hardware is close but not comfortable, read the Local LLMs hub next. If you are shopping the next GPU or mini PC, use Homelab Gear for the rig I actually run.
Try it, then go deeper
Open the free local LLM hardware checker, pick your GPU and RAM, and see what fits. When you are ready to buy or rebalance the lab, Homelab Gear has the hardware side.
FAQ
Is the local LLM hardware checker free?
Yes. It runs in the browser with no signup.
Are the speed numbers exact?
No. They are community-style estimates. Real tokens/sec depend on quantization, context length, and the runtime you use.
What should I read after the checker?
Start with the Local LLMs hub. For hardware shopping, open Homelab Gear.