kapynOpen Source

local-enough , Measure whether a back-office AI task can run on your own hardware: accuracy vs gold, p50/p95 latency, $/1k tasks, amort

Local-enough is an open-source Python tool for benchmarking back-office AI tasks against local hardware. It measures accuracy against gold standards, latency percentiles, and amortized hardware costs to find break-even volumes. Developers can use its OpenAI-compatible router to dynamically route each task to the cheapest model that meets a specific quality threshold.

GitHub·Sep 29, 2026

Opening Kapyn…