📝 docs(readme): drop Quire Cloud subsection

The Privacy section in the README advertised a planned paid hosted
service that isn't part of this repository. Out of scope for the OSS
README. Removes the subsection and collapses the now-redundant
self-hosted/Cloud h3 split back into a flat Privacy section.

Author: vitofico

README.md

@@ -70,11 +70,7 @@ is Quire-specific.
 
 ## Privacy
 
-Self-hosted Quire — what this repository ships — is the only Quire that
-exists today. It stays Apache-2.0, calibre-web-only, and **no telemetry,
-ever.**
-
-### Self-hosted (this repository)
+Apache-2.0, calibre-web-only, **no telemetry, ever.**
 
 - No analytics, no crash reporting, no third-party SDKs.
 - Network calls go to your calibre-web instance and your Quire Server.
@@ -88,30 +84,6 @@ ever.**
 - Credentials are stored in Android Keystore (hardware-backed where the
   device supports it).
 
-### Quire Cloud (future paid hosted service, separate)
-
-Quire Cloud is a planned, separate, paid hosted service operated by the
-author. It does not exist yet and is not part of this repository.
-
-> Quire Cloud is a paid service that processes service data only to
-> provide its features. See the privacy notice for what we store, why,
-> for how long, and the subprocessor list.
-
-Before any Cloud private beta opens, the following will ship as
-non-negotiable prerequisites:
-
-- A published Cloud privacy notice covering data categories, processing
-  purposes, retention, user rights, and the full subprocessor list.
-- A data-subject access / export endpoint ("download everything").
-- Cascading account deletion (delete account → all associated rows go
-  with it).
-
-None of the above changes anything about the self-hosted edition.
-There is no Cloud sign-in requirement in the OSS app, no telemetry
-back-channel, and no feature gated behind Cloud. Self-hosters and
-Cloud users will hit the same `quire_server` codebase via the same
-HTTP API; only the operator and the data-controller differ.
-
 ## AI features (optional)
 
 Quire optionally calls AI for book insights and library analysis. AI is