Lucinity is Poor to agents.

What we found

• —An agent can reach a complete, single-page API reference (Redoc-generated, v3.3) covering Actors, Accounts, Transactions, Observations, Cases, Webhooks, Reporting and Segmentation — but discovery is weak: no llms.txt, no llms-full.txt, no .well-known/mcp.json, and no downloadable OpenAPI file at any standard path, so an agent must scrape a 438KB HTML page rather than fetch a machine-readable spec.
• —The docs.lucinity.com sitemap lists /docs/* documentation URLs that now 404 — the live reference has moved to the site root — so an agent following the sitemap lands on dead pages, a stale-discovery trap.
• —Comprehension is the strong point: endpoint summaries are task-oriented (e.g. 'Creates a new actor entity and assigns it to the specified case'), a numbered Quickstart walks through the core ingest-then-initialize-batch flow, and Redoc gives clean H1→H2→H3 nesting with method/URL shown answer-first.
• —Capability boundaries are stated only inline and unevenly — e.g. 'blocking=true is only supported and MUST be set', 'Transaction screening ... may not be configured for your environment', limit 'not greater than 1000' — with no consolidated limitations page and no documented rate limits, so an agent discovers many limits only by failing.
• —Access is gated on humans: API keys are obtained by contacting 'your Lucinity Customer Success Team', and there is no self-serve sandbox, no public OpenAPI download, and no MCP-registry presence, so an autonomous agent cannot go from docs to a working call unaided.

What to change

Prioritized by impact on discoverability. You (or your docs platform) deploy these — Discry never touches your API.

1. 01Publish the underlying OpenAPI spec at a stable public path (e.g. https://docs.lucinity.com/openapi.json). The Redoc docs are already generated from a v3.3 spec — exposing the raw file is the single highest-impact discovery fix (weight 5) and lets agents load endpoints machine-readably instead of scraping 438KB of HTML.
2. 02Add an llms.txt at docs.lucinity.com describing the API (Actors, Accounts, Transactions, Observations, Cases, Webhooks) and linking the Quickstart, Authentication, and OpenAPI spec. This is table-stakes discovery and would give agents a <5K-token entry point into an otherwise monolithic page.
3. 03Fix the stale sitemap: the listed /docs/* URLs 404. Regenerate it to point at the live section anchors so agents and crawlers reach real content.
4. 04Turn error documentation into recovery guidance: for the common 400/401/404/422/500 responses, state which field or condition triggered it and what to do (e.g. how to correct a 422 'required fields missing', how to re-auth on 401), and document rate limits and any Retry-After behaviour.
5. 05Consolidate capability boundaries into one place — supported screening operations, environment-gated features, pagination caps, and rate limits — so an agent can plan around them instead of discovering them through failed calls.