Decision tools

Use a tool when you need an answer from your own inputs. Start with hardware fit unless you already know the exact model and app you want.

Start with a concrete decision

Local LLM Guide tools help you make practical decisions before downloading a model or committing to an app.

Start with the RAM and VRAM calculator if you already have a model in mind. Enter the model size, model compression, conversation length, and memory available on your computer. On a Windows PC, use dedicated GPU memory rather than shared-memory figures. On an Apple Silicon Mac, leave room for macOS and other apps in the shared memory pool.

Use the local setup recommender when you know your computer and task but have not chosen an app. A desktop chat app and a browser workspace for documents can require different setup paths. Use App comparison when you want to examine those differences side by side.

Use the Cloud vs local cost calculator when deciding whether your usage justifies a hardware purchase. Include electricity, any new computer cost, and your expected cloud usage. Reuse hardware you already own as a separate scenario so you can see which assumption changes the estimate.

Use the Token and PDF counter before pasting a long source into a chat. It reads text in your browser and helps reserve room for instructions, chat history, and the reply. Its selected tokenizer may differ from your model's tokenizer. Scanned PDFs need text recognition before extracted text can be counted.

Use Browser AI chat to try a small downloaded model on a supported browser. Read the model size and browser requirements before downloading. A browser model still uses your computer's memory and can give incorrect answers; begin with a short task whose answer you can check.