01
Discover
Choose your workload, model artifact, context target, usable memory, concurrency and privacy needs.
Free interactive decision tool · built with Base44
Answer the practical infrastructure questions together: how much memory might your AI workload need, can it fit the hardware you already own, and should you start with local, VPS, GPU cloud or hybrid?
This is a planning estimate with documented assumptions, not a live GPU quote or an independent benchmark. Only variants from AIG's published measured runs appear as audited presets; other models can be entered manually. Measured performance applies only to the documented test conditions.
01
Choose your workload, model artifact, context target, usable memory, concurrency and privacy needs.
02
Estimate planning memory with visible reserves. The tool identifies headroom, but does not claim measured runtime compatibility.
03
See a rule-based starting architecture and viable alternatives across local, VPS, GPU cloud and hybrid infrastructure.
04
See which information is missing before buying or deploying, then compare AIG's other calculators and measured benchmarks.
After using Planner Lite, compare the result with our specialised calculators and measured local-model data. The $19 Recommendation Kit offers editable documents for the full decision workflow; paid Planner software remains in development.