d3ro-voice/apps
Yun Chan 648c7e4f4d refactor(meeting): share the transcript polish prompt with tests
Move the polish instructions into their own module so the prompt the app
sends, the unit tests and the local-model probe use the same text. Tests
now pin the two rules that matter: fix a misheard name or in-house term to
the learned spelling, and never swap an ordinary word on similarity alone.

Verified against the local model with the real learned corpus: 하람록에 was
rewritten to HaramLog에 and the dictionary pronunciation 디에로 to D3RO,
while a natural sentence with no learned term came back with no word
substitutions. Recorded the trap that this model returns empty content
unless thinking is disabled.
2026-11-06 21:47:16 +09:00
..
admin chore(release): v1.16.4 2026-11-06 17:23:22 +09:00
admin-swagger chore(release): v1.16.4 2026-11-06 17:23:22 +09:00
api-server chore(release): v1.16.4 2026-11-06 17:23:22 +09:00
api-server.Tests feat(release): prepare 1.1.0 candidate 2026-08-29 18:33:45 +09:00
desktop refactor(meeting): share the transcript polish prompt with tests 2026-11-06 21:47:16 +09:00
mobile-rn chore(release): v1.16.4 2026-11-06 17:23:22 +09:00
web chore(release): v1.16.4 2026-11-06 17:23:22 +09:00