Phase 4 구현: Ollama LLM 연동 (다듬기, 번역, 스트리밍)

- LocalLLMService: Ollama REST API, 스트리밍 NDJSON 파싱, 5초 가용성 폴링
- 시스템 프롬프트: refine/translate/summarize/grammar/expand/custom 6개 액션
- VoiceModeService: 전사→LLM 후처리→텍스트 삽입, LLM 실패 시 원본 폴백
- LLM IPC 핸들러: status/models/process/cancel/serverUrl 8개
- StatusBar: Ollama 연결 상태 + 활성 모델 표시
- Preload: llm API 섹션 추가
- Bootstrap: llm-polling 초기화 단계 추가
This commit is contained in:
Yun Chan 2026-04-05 02:18:26 +09:00
parent 517210af2f
commit 4a5cf6c819
10 changed files with 663 additions and 9 deletions

View file

@ -25,3 +25,4 @@ export { getHotkeyService } from './HotkeyService'
export { getVoiceModeService } from './VoiceModeService'
export { getAudioCaptureService } from './AudioCaptureService'
export { getTextInsertService } from './TextInsertService'
export { getLocalLLMService } from './LocalLLMService'