Local AI Compatibility Radar
Probe your browser's WebGPU capabilities and run local AI readiness checks. Learn if your device can execute small language models entirely in browser.
GPU & RAM Setup
Run the radar.
Actual token rates depend on context bounds, GPU cache, and driver state.
Use estimates for initial planning checks.
Model Fit Matrix
| Tier size | Examples | Estimate Fit | Description |
|---|
Copyable Compatibility Report
Local Performance Log
Log your actual local benchmark speeds below. Records are stored locally in your browser (no server sync). All entries represent local logs, not guaranteed metrics.
| Hardware Model | Model Size | Quant | Runtime | Local Log Speed | Context Window | Action |
|---|
Planning Boundaries
Ready to run a real browser-local prompt test?
Use the Local Prompt Tester to check WebGPU support, manually load a small WebLLM model, stream a response, and capture local speed metrics. The model is not downloaded until the user clicks the load button.
Browser-local inference is device-dependent. Avoid absolute privacy or quality claims; use measured local results and clear limits.
What this tool does
- Probes browser context dynamically for active WebGPU API support.
- Estimates local execution suitability (fit scores) for small language models (SLMs).
- Maps model tiers (from 1B to 70B+ weights) to hardware capabilities like RAM and VRAM.
- Checks device readiness for local tools like LM Studio, Ollama, and WebLLM.
Best for
Developers, AI builders, and researchers assessing browser compatibility and client-side hardware suitability for offline local LLM operations.
Privacy Notes & Verified Badges
No hardware specifications or IP details are uploaded. All capability checks run fully locally via client API queries.
Example Workflow
- Open the checker in the browser you plan to use.
- Run the compatibility test.
- Review WebGPU, browser, and device support signals.
- Read any warnings or limitations.
- Compare results with local AI tool requirements.
- Use the result as a planning guide, not a guarantee.
Common Use Cases
Determine if your browser's security boundaries and flags support WebGPU canvas allocation.
Map RAM benchmarks and quantization tolerances to predict local model loading boundaries.
Run dynamic hardware probes to check if the active GPU driver exposes compute pipelines.
Provide a structured device suitability passport to explain why local SLM loading fails on legacy hardware.
Frequently Asked Questions (FAQ)
What does this checker test?
It probes your browser for WebGPU support, examines available system memory, and calculates compatibility tiers for running small local AI language models (like Llama 3 or Phi 3) client-side.
Does WebGPU support guarantee local AI will run fast?
No. WebGPU support allows the browser to access your graphics card (GPU) for hardware acceleration. The actual execution speed depends on your GPU class, VRAM capacity, and model quantization.
Which browsers work best for WebGPU?
Modern Chromium-based browsers (Chrome, Edge, Opera) and Firefox (on supported platforms) offer the most stable WebGPU implementations. Safari support is evolving; last checked June 14, 2026. Confirm current browser release notes before relying on production compatibility.
Does this page run a model or upload my data?
No. This is a lightweight compatibility radar that checks hardware capabilities and API availability. It does not download heavy models or send device specs to our servers.