feat: add Hermes WebUI provisioning and English Dragon HD Omni voices

tests/README.md

@@ -4,13 +4,26 @@
 
 ```sh
 python -m venv .venv-tests
-.venv-tests/bin/python -m pip install pytest fastapi httpx
+.venv-tests/bin/python -m pip install pytest pytest-asyncio -r control/requirements.txt
 .venv-tests/bin/python -m pytest --noconftest tests/unit/test_accounts.py tests/unit/test_provisioner.py tests/unit/test_token_bootstrap.py tests/unit/test_deploy_rollback.py -q
 agent/.venv/bin/python -m unittest discover -s tests/unit -p test_hermes_llm.py -v
 ```
 
 Covers authorization, revocation, CSRF, lifecycle/restart failures, Docker ownership, authenticated server-derived room routing, SSE framing, content-only output, cancellation, redacted errors, and wait/final-answer callbacks.
 
+## Azure English voice regression
+
+```sh
+.venv-tests/bin/python -m pytest --noconftest tests/unit/test_voice_catalog.py tests/unit/test_voice_settings_ui.py -q
+```
+
+Catalog tests mock Azure and the authenticated account boundary. Browser tests
+require Playwright Chromium and use local assets with mocked APIs, never the NAS.
+They cover the Ava Omni default, legacy HD preference migration, English accents,
+search/selection, cache/failure recovery, and anonymous/revoked-account denial.
+Derived Omni personas follow Microsoft's documented naming convention; individual
+voice synthesis and regional availability still require live Azure qualification.
+
 ## Legacy tests
 
 Top-level `test_*.py` and `conftest.py` describe the former OpenWebUI/MCP architecture and default to live Unraid. **Do not run them as the migrated regression suite.** Retained only as historical references: weather/dispatch/local memory and anonymous-token assumptions no longer apply. Hope Music remains an independent deployment.