benchouse-data

Community-contributed capability profiles for the analytics agents and context/semantic layers compared at benchouse.ai (/agents and /layers). This is the only place those facts live — benchouse.ai is a static-rendered read of this repo, nothing more.

Every profile is facts only — there is deliberately no rating, color, or opinion field, so the comparison stays neutral even though it's crowd-sourced. See schema/agent.md and schema/layer.md for the field reference, and CONTRIBUTING.md for how to submit a fact via PR.

Layout

agents/<slug>.json       one file per analytics agent
layers/<slug>.json       one file per context/semantic layer
schema/                  field reference for contributors

A profile's slug is soft-linked to benchouse.ai's leaderboard by model.Slugify of the entry's agent (agents) or semanticLayer (layers) field — a profile doesn't require a benchmarked entry, and a benchmarked entry doesn't require a profile.