Skip to content

No install, no server

Run AI models in your browser.

Live demos one click away, download sizes measured to the byte. The fit math for every local model sits right below.

In-browser
54
Local models
163
Devices
40
Free
Always
Browser check LIVE

Checking for WebGPU…

Nothing downloads until you press Run on the model's page.

Covers
Text generationImage modelsVideo modelsAudio modelsIn the browserApple SiliconNVIDIA / AMDText generationImage modelsVideo modelsAudio modelsIn the browserApple SiliconNVIDIA / AMD
Runs with
Transformers.jsOllamaLM Studiollama.cppMLXWebLLMvLLMJanKoboldCppGPT4AllTransformers.jsOllamaLM Studiollama.cppMLXWebLLMvLLMJanKoboldCppGPT4All

The reference layer

Can I run any AI model locally?

Pick a model and your device. Get a yes, tight, or no, plus the command to run it. 163 models, 40 devices, every size measured.

Yes, it runsfast
needs 6.4 GBusable 10.5 GB
4k

Runs at Q4_K_M using ~6.4 GB of ~10.5 GB usable. You have room for Q8_0 for higher quality.

See the full breakdown
$ollama run llama3.1:8b

Best on macOS: LM Studio · Q4_K_M recommended

How it works

  1. 01 · Model
    Size × quant
  2. 02 · Memory math
    Weights + KV cache + overhead
  3. 03 · Verdict
    Yes · Tight · No

Every verdict comes from this math, sourced and documented. See the full methodology →

Popular models

All 163 models →

Find your model another way

Why the numbers are reliable

1

Real sizes

Quant sizes from HuggingFace GGUF repos, Ollama and vendor specs. Every number sourced, never guessed.

2

Honest memory math

KV cache and overhead included, on realistic usable memory (Apple unified ~66-75%, GPU VRAM minus driver).

3

The right tool

A different winner per platform: MLX on Mac, CUDA on Windows, vLLM on Linux, PocketPal on phones.

The extremes

Check any pair →

The same memory math, at the edges of the catalog.

Frequently asked

Can I run AI models in my browser without installing anything?

Yes. 54 models in the catalog run entirely in the tab with Transformers.js, on the GPU over WebGPU (Chrome, Edge, Safari 26+) with an automatic WebAssembly fallback elsewhere. Whisper Tiny transcribes speech after a 72.6 MB download, SmolLM2-135M-Instruct chats after 112.2 MB, Kokoro-82M speaks after 147.4 MB. Six pages run them live with one click; every download size is measured from the HuggingFace file tree, not estimated.

How do I know if my computer can run a local AI model?

Compare the model's memory needs to your usable memory. A 7-8B text model at Q4_K_M needs about 6-7 GB, so it runs on a 16 GB Mac or a 12 GB GPU. Image and video diffusion models typically need 4-12 GB of GPU or Apple Silicon VRAM. Audio models (Whisper, Kokoro) run on CPU and need 1-4 GB. localmodel.run does this math for 163 models across 40 devices.

Can I run AI models locally on a Mac?

Yes. Apple Silicon shares unified memory, so a 16 GB Mac runs 7-8B models and a 64 GB+ Mac runs 70B. Use LM Studio (which ships MLX) for a GUI, or mlx-lm for the most speed. vLLM is not a Mac tool, it is a Linux/CUDA serving engine.

Can I run AI models on my phone?

Yes, within limits. iPhones and Android flagships realistically run 1B-4B text models. For text: PocketPal AI works on both iOS and Android; Apple Foundation Models is built into iOS 26. For images on iPhone: Draw Things supports diffusion models locally. For audio: Whisper (speech-to-text) runs on both iOS and Android.

Which is the best tool to run models locally?

On Mac, start with LM Studio, it ships MLX and has a GUI. On Linux, Ollama for quick chat, vLLM if you are serving traffic. On phones: for text, PocketPal AI (iOS and Android) or Apple Foundation Models (iOS 26); for images on iPhone, Draw Things; for audio, whisper.cpp. Each device page links the right tool so you do not have to guess.

Estimates, not guarantees. See how we calculate and our sources.