CLOVA Speech Lecture MCP Server

This is an MCP server for lecture recording STT and summarization.
It calls Naver CLOVA Speech / CLOVA Studio APIs from the Google ADK agent.

Provided Tools

Related MCP server: audio-transcriber

Getting Started

1. API Key Issuance

CLOVA Speech (STT)

1. Log in to Naver Cloud Platform

2. AI Services → CLOVA Speech → Apply for use

3. Copy the Secret Key after creating the service

CLOVA Studio (Summarization)

1. AI Services → CLOVA Studio → Apply for use

2. Create an app in the Playground

3. Copy the API key and App ID

2. .env Configuration

cp .env .env.local   # 실제 값으로 수정

CLOVA_SPEECH_API_KEY=실제_키_입력
CLOVA_STUDIO_CLIENT_ID=실제_Client_ID
CLOVA_STUDIO_CLIENT_SECRET=실제_Client_Secret
CLOVA_STUDIO_APP_ID=실제_앱_ID

3. Execution

# Docker
docker-compose up -d

# 로컬 실행
pip install -r requirements.txt
python app/main.py

4. ADK Agent Connection

# agent.py 예시
tools = [
    MCPToolset(
        connection_params=SseServerParams(url="http://localhost:8002/sse")
    )
]

Lecture Processing Flow

[강의 오디오 파일]
       ↓
transcribe_lecture_submit(file_path, enable_diarization=True)
       ↓ (task_id 반환)
get_transcription_result(task_id)   ← 완료까지 자동 폴링
       ↓ (full_text 반환)
summarize_lecture(text)
       ↓
[요약문 + 키워드]

Supported Audio Formats

.wav .mp3 .flac .m4a .aac .ogg

Supported Languages