How It Works
Four steps, zero guesswork. The same deterministic engine powers every check on the site.
- 01
Enter your laptop specifications
CPU, RAM, GPU, VRAM, storage, OS, architecture and NPU — or load a saved profile.
- 02
Choose what you want to run
A local LLM, a coding agent, an image generator, an IDE — anything in the database.
- 03
The rules engine checks every requirement
Explicit IF/THEN rules with priorities: CRITICAL, HIGH, MEDIUM, LOW, OPTIONAL.
- 04
Get a clear YES, LIMITED or NO
Plus a full breakdown, the reason for every failure, and upgrade suggestions.
Popular checks
Ollama
Ollama is the most popular local LLM runner. It downloads and serves open models (Llama, Qwen, Mistral, Gemma and more) with a simple CLI and local API. Actual model performance depends on which model you run — check the LLM size classes in this database.
Min: 8 GB RAM · shared GPU OK · 10 GB storage
LM Studio
LM Studio is a polished desktop app for discovering, downloading and chatting with local LLMs, including an OpenAI-compatible local server. It supports GGUF models on CPU and GPU and offers per-model compatibility hints at download time.
Min: 8 GB RAM · shared GPU OK · 10 GB storage
GPT4All
GPT4All by Nomic AI is an open-source desktop app for running local LLMs on consumer hardware, including CPUs without dedicated GPUs. It includes a local docs chat feature and emphasises privacy-first, offline operation.
Min: 8 GB RAM · shared GPU OK · 8 GB storage
Jan
Jan is an open-source ChatGPT alternative that runs 100% offline on your machine. It uses llama.cpp under the hood, supports GGUF models and exposes an OpenAI-compatible local API server.
Min: 8 GB RAM · shared GPU OK · 8 GB storage
llama.cpp
llama.cpp is the foundational C/C++ inference engine for GGUF models. It runs on almost anything — from old laptops to workstations — with optional CUDA, Metal, Vulkan and OpenCL acceleration. It is a command-line tool, so expect manual compilation or community binaries.
Min: 4 GB RAM · shared GPU OK · 2 GB storage
Open WebUI
Open WebUI is a self-hosted web interface for local LLM runners (Ollama, llama.cpp and OpenAI-compatible endpoints). It adds multi-user chat, RAG document uploads and model management, typically deployed via Docker.
Min: 8 GB RAM · shared GPU OK · 5 GB storage
Why AgentCheck?
No AI decisions
Every verdict comes from explicit IF/THEN rules over stored requirements — same input, same result, always.
No API dependency
The engine runs in your browser with a bundled database. It works even with every external service off.
Transparent rules
Rule priorities, thresholds and the decision tree are published on the How It Works page.
Detailed reports
Per-requirement pass/fail, dual minimum/recommended verdicts, score, and upgrade suggestions.
Free basic checking
No account needed. Profiles stay in your browser via localStorage. Nothing is uploaded.
ai laptop requirements
AI Laptop Requirements in 2026: The Complete Guide
What CPU, RAM, GPU, VRAM and storage do you actually need for local AI in 2026? A structured, no-hype guide with real numbers for agents, local LLMs and image models.
9 min read · updated 2026-09-01
ram guides
How Much RAM Do You Need for Local AI? 8 vs 16 vs 32 vs 64 GB
RAM is the number-one bottleneck for local LLMs. Here is exactly what each memory tier can and cannot run, with the arithmetic behind the numbers.
7 min read · updated 2026-09-01
vram guides
How Much VRAM Do AI Agents Need? A Practical Breakdown
AI agents themselves are light — their local LLMs are not. A component-by-component VRAM breakdown for agent stacks, from chat to code execution to browser automation.
6 min read · updated 2026-08-30