Our current scope
The first research edition contains 97 product, tool, framework, and feature records; 464 selected documentary capability entries; and 179 proposed task areas across 20 categories. It is a selective research set, not a complete map of every product or feature.
Three questions, kept separate
Is it documented? A direct source attributes a capability to a named product surface. Does it work in our configuration? A recorded runtime test would establish observed behavior. How well does it compare? A matched evaluation would support a narrower comparative finding. Current catalog entries answer the first question.
A general assistant can span categories
A system can appear in more than one category when a recorded capability supports that kind of work. Category placement is not evidence that it supports every task in the category. We keep task definitions separate from claim records until an explicit evidence mapping is reviewed.
What a future score must include
A numerical performance score requires a versioned task protocol, fixed inputs, disclosed product and model configuration, run dates, observed outputs, and a stated rubric. Repeated runs and independent judgments must be used where the task needs them. Cost, latency, output quality, and workflow usability should remain visible as distinct dimensions.
How updates should work
A change to a vendor source should enter an editorial queue. A researcher checks the exact claim, route, plan, and lifecycle status before publication. A release announcement does not automatically establish availability in every account. Source review dates change only when the claim is reviewed; they are not refreshed just because the site is rebuilt.
Research status and limitations
No runtime evaluations or numerical performance rankings have been published. Some records have partial evidence and are labeled accordingly. Individual product profiles remain research previews and are excluded from search indexing until editorial review. Category guides, curated comparisons, and decision guides provide the published starting points for this research edition.