Local utility · Last updated: 2026-06-10
Local AI Atlas Guide
Learn how to safely profile your local hardware and match it with open-source local AI model tiers 100% locally.
Tool shortcut includedNo data uploadedWebGPU check
Short Answer: The Local AI Atlas provides a dynamic compatibility lookup to match user hardware specs (RAM, VRAM, WebGPU) with structured model tiers (Lite, Story, Desktop, Power) 100% locally in the browser.
When to use this
Use this hardware check workflow when deciding which open local models (like Llama 3, Phi 3, or Gemma) to download for LM Studio, Ollama, or in-browser WebLLM frameworks, without manually calculating VRAM overhead.
Basic workflow
- Trigger the WebGPU signal probe to check browser runtime capabilities.
- Input your system specs: System RAM, dedicated GPU VRAM, and target task goal.
- Review the compatibility score dashboard showing fits from Lite to Power tiers.
- Copy your compatibility passport to share or reference during local setup.
Quality and safety notes
- VRAM limits are hard constraints. Running a model that exceeds VRAM will cause heavy fallback slowdowns or memory crashes.
- Browser memory allocations are typically restricted compared to desktop runtimes (like Ollama or LM Studio).
- Dataset parameters use starter benchmarks and should be cross-checked with official model cards before deploying production apps.
Frequently Asked Questions
How does the compatibility scoring work?It checks your hardware specs (RAM, VRAM) and browser capabilities (WebGPU) against parameters required by model tiers, generating a compatibility score out of 100.
Does this tool collect or upload my hardware profile?No. The hardware profiling, probe results, and compatibility matching occur entirely client-side. No data is sent to any server.