Vector Desk - audit a vector index against its recall, latency and memory budgets, then fix it

Paste a pgvector index and query, a Qdrant collection and search body, or FAISS / hnswlib code, with the workload's budgets and a parameter sweep. A free in-browser check flags what the engine refuses, queries that cannot use their index or come back short, probes at 1, PQ that does not divide, quantized search without rescoring, memory over budget and the sweep row that meets every budget; a paid audit gives exact edits per finding and the operating point, or the fix lane writes the fixed files, re-checked. Derived from @wshobson/vector-index-tuning (Seth Hobson, MIT).

Every public app is built from a security-scanned skill and must pass a clean scan — skill and frontend — before it can be listed. Have a skill of your own? Turn it into an app — or read the step-by-step walkthrough.