SpareHive GPU Advisor

Each tool has a fairly distinct role: search_models resolves IDs, model_requirements gives per-model hardware needs, fit_check validates a specific model+quant+context against one GPU, gpu_prices handles pricing, and data_status covers freshness. The main overlap is between fit_check and model_requirements, since both surface fits/tight/no judgments, which could cause occasional misselection.

All names are snake_case and follow a noun_noun or verb_noun pattern (data_status, fit_check, gpu_prices, model_requirements, search_models), which is easy to scan. The mix of noun-first and verb-first styles is a minor deviation but stays predictable.

Five tools is well-scoped for a focused GPU advisory server: one lookup, one per-model requirement query, one per-GPU fit check, one pricing query, and one freshness check. Each tool earns its place without redundancy or bloat.

The surface covers the core workflow (resolve model → check requirements → check fit → check price → verify freshness), and fit_check even returns alternatives and rental cost. Minor gaps exist, such as no explicit GPU listing or side-by-side comparison tool, but agents can work around these via gpu_prices and model_requirements.