Every local AI tool, by category

This page lists every category on LaptopMind and every tool in it. The categories follow what people actually set out to do when they run AI on their own computer, so each tool appears once, under the job it is mainly used for, even when it does more.

Desktop apps are the easiest way in: install, pick a model, start a chat. Local model runners do the same work from the command line and usually start a server that other apps can share. Inference engines and runtimes sit one level lower, squeezing more speed out of a particular kind of hardware, and they matter most when you serve a model to several people or build your own app.

Local image generation covers the interfaces for open diffusion models, from simple prompt boxes to node graphs. On-device speech and voice covers transcription and text-to-speech that never upload your audio. Private document chat is for asking questions about your own files with a local model, and local coding assistants bring completion and chat into your editor without sending code away.

Three categories support the rest. Model hubs and formats explains where open weights come from and the file types you will meet. Fine-tuning and quantization covers adapting a model to your own data and shrinking it to fit your memory. Hardware and benchmarking utilities are the small tools that show if your GPU is the bottleneck.

The lines between categories are not strict. A desktop app may include a local server, a document chat tool may run its own model, and an image interface may also generate video. When a tool covers more than its main job, its page says so, and each category page links to the related categories so you can compare across the line.

If you are not sure where to start, begin with a desktop app and a small model. Once you know what you want from it, the other categories make more sense, and each category page explains what to check before you commit.