diff --git a/app/utils/prompts/prompts.py b/app/utils/prompts/prompts.py
index 6f7a70e..d0cbc09 100644
--- a/app/utils/prompts/prompts.py
+++ b/app/utils/prompts/prompts.py
@@ -4,69 +4,73 @@ from config import prompt_settings
 from app.utils.logger import get_logger
 from app.utils.prompts.schemas import *
 from functools import lru_cache
+import openpyxl
 
 logger = get_logger("prompt")
 
 class Prompt():
-    prompt_template_path : str #프롬프트 경로
-    prompt_template : str # fstring 포맷
-    prompt_model : str 
-    
-    prompt_input_class = BaseModel # pydantic class 자체를(instance 아님) 변수로 가짐
+    sheet_name: str
+    prompt_template: str
+    prompt_model: str
+
+    prompt_input_class = BaseModel
     prompt_output_class = BaseModel
-              
-    def __init__(self, prompt_template_path, prompt_input_class, prompt_output_class, prompt_model):
-        self.prompt_template_path = prompt_template_path
+
+    def __init__(self, sheet_name, prompt_input_class, prompt_output_class):
+        self.sheet_name = sheet_name
         self.prompt_input_class = prompt_input_class
         self.prompt_output_class = prompt_output_class
-        self.prompt_template = self.read_prompt()
-        self.prompt_model = prompt_model
+        self.prompt_template, self.prompt_model = self._read_from_excel()
+
+    def _read_from_excel(self) -> tuple[str, str]:
+        wb = openpyxl.load_workbook(prompt_settings.PROMPT_EXCEL_FILE, read_only=True)
+        try:
+            ws = wb[self.sheet_name]
+            data = {}
+            for row in ws.iter_rows(min_row=2, values_only=True):
+                key, value = row[0], row[1]
+                if key and value:
+                    data[key] = value
+        finally:
+            wb.close()
+
+        return data["input"], data["model"]
 
     def _reload_prompt(self):
-        self.prompt_template = self.read_prompt()
+        self.prompt_template, self.prompt_model = self._read_from_excel()
 
-    def read_prompt(self) -> tuple[str, dict]:
-        with open(self.prompt_template_path, "r") as fp:
-            prompt_template = fp.read()
-
-        return prompt_template
-
-    def build_prompt(self, input_data:dict) -> str:
+    def build_prompt(self, input_data: dict) -> str:
         verified_input = self.prompt_input_class(**input_data)
-        build_template = self.prompt_template
-        build_template = build_template.format(**verified_input.model_dump())
+        build_template = self.prompt_template.format(**verified_input.model_dump())
         logger.debug(f"build_template: {build_template}")
         logger.debug(f"input_data: {input_data}")
         return build_template
-    
+
 marketing_prompt = Prompt(
-    prompt_template_path = os.path.join(prompt_settings.PROMPT_FOLDER_ROOT, prompt_settings.MARKETING_PROMPT_FILE_NAME),
-    prompt_input_class = MarketingPromptInput,
-    prompt_output_class = MarketingPromptOutput,
-    prompt_model = prompt_settings.MARKETING_PROMPT_MODEL
+    sheet_name="marketing",
+    prompt_input_class=MarketingPromptInput,
+    prompt_output_class=MarketingPromptOutput,
 )
 
 lyric_prompt = Prompt(
-    prompt_template_path=os.path.join(prompt_settings.PROMPT_FOLDER_ROOT, prompt_settings.LYRIC_PROMPT_FILE_NAME),
-    prompt_input_class = LyricPromptInput,
-    prompt_output_class = LyricPromptOutput,
-    prompt_model = prompt_settings.LYRIC_PROMPT_MODEL
+    sheet_name="lyric",
+    prompt_input_class=LyricPromptInput,
+    prompt_output_class=LyricPromptOutput,
 )
 
 yt_upload_prompt = Prompt(
-    prompt_template_path=os.path.join(prompt_settings.PROMPT_FOLDER_ROOT, prompt_settings.YOUTUBE_PROMPT_FILE_NAME),
-    prompt_input_class = YTUploadPromptInput,
-    prompt_output_class = YTUploadPromptOutput,
-    prompt_model = prompt_settings.YOUTUBE_PROMPT_MODEL
+    sheet_name="yt_upload",
+    prompt_input_class=YTUploadPromptInput,
+    prompt_output_class=YTUploadPromptOutput,
 )
 
 @lru_cache()
-def create_dynamic_subtitle_prompt(length : int) -> Prompt:
-    prompt_template_path=os.path.join(prompt_settings.PROMPT_FOLDER_ROOT, prompt_settings.SUBTITLE_PROMPT_FILE_NAME)
-    prompt_input_class = SubtitlePromptInput
-    prompt_output_class = SubtitlePromptOutput[length]
-    prompt_model = prompt_settings.SUBTITLE_PROMPT_MODEL
-    return Prompt(prompt_template_path, prompt_input_class, prompt_output_class, prompt_model)
+def create_dynamic_subtitle_prompt(length: int) -> Prompt:
+    return Prompt(
+        sheet_name="subtitle",
+        prompt_input_class=SubtitlePromptInput,
+        prompt_output_class=SubtitlePromptOutput[length],
+    )
 
 
 def reload_all_prompt():
diff --git a/app/utils/prompts/templates/20260129_marketing_prompt.txt b/app/utils/prompts/templates/20260129_marketing_prompt.txt
deleted file mode 100644
index 2bf9307..0000000
--- a/app/utils/prompts/templates/20260129_marketing_prompt.txt
+++ /dev/null
@@ -1,64 +0,0 @@
-
-[Role & Objective]
-Act as a content marketing expert with strong domain knowledge in the Korean pension / stay-accommodation industry.
-Your goal is to produce a Marketing Intelligence Report that will be shown to accommodation owners BEFORE any content is generated.
-The report must clearly explain what makes the property sellable, marketable, and scalable through content.
-
-[INPUT]
-- Business Name: {customer_name}
-- Region: {region}
-- Region Details: {detail_region_info}
-
-[Core Analysis Requirements]
-Analyze the property based on:
-Location, concept, and nearby environment
-Target customer behavior and reservation decision factors
-Include:
-- Target customer segments & personas
-- Unique Selling Propositions (USPs)
-- Competitive landscape (direct & indirect competitors)
-- Market positioning
-
-[Key Selling Point Structuring – UI Optimized]
-From the analysis above, extract the main Key Selling Points using the structure below.
-Rules:
-Focus only on factors that directly influence booking decisions
-Each selling point must be concise and visually scannable
-Language must be reusable for ads, short-form videos, and listing headlines
-Avoid full sentences in descriptions; use short selling phrases
-Do not provide in report
-
-Output format:
-[Category]
-(Tag keyword – 5~8 words, noun-based, UI oval-style)
-One-line selling phrase (not a full sentence)
-Limit:
-5 to 8 Key Selling Points only
-Do not provide in report
-
-[Content & Automation Readiness Check]
-Ensure that:
-Each tag keyword can directly map to a content theme
-Each selling phrase can be used as:
-- Video hook
-- Image headline
-- Ad copy snippet
-
-
-[Tag Generation Rules]
-- Tags must include **only core keywords that can be directly used for viral video song lyrics**
-- Each tag should be selected with **search discovery + emotional resonance + reservation conversion** in mind
-- The number of tags must be **exactly 5**
-- Tags must be **nouns or short keyword phrases**; full sentences are strictly prohibited
-- The following categories must be **balanced and all represented**:
-  1) **Location / Local context** (region name, neighborhood, travel context)
-  2) **Accommodation positioning** (emotional stay, private stay, boutique stay, etc.)
-  3) **Emotion / Experience** (healing, rest, one-day escape, memory, etc.)
-  4) **SNS / Viral signals** (Instagram vibes, picture-perfect day, aesthetic travel, etc.)
-  5) **Travel & booking intent** (travel, getaway, stay, relaxation, etc.)
-
-- If a brand name exists, **at least one tag must include the brand name or a brand-specific expression**
-- Avoid overly generic keywords (e.g., “hotel”, “travel” alone); **prioritize distinctive, differentiating phrases**
-- The final output must strictly follow the JSON format below, with no additional text
-
-  "tags": ["Tag1", "Tag2", "Tag3", "Tag4", "Tag5"]
diff --git a/app/utils/prompts/templates/lyric_prompt.txt b/app/utils/prompts/templates/lyric_prompt.txt
deleted file mode 100644
index 5b04be0..0000000
--- a/app/utils/prompts/templates/lyric_prompt.txt
+++ /dev/null
@@ -1,77 +0,0 @@
-
-[ROLE]
-You are a content marketing expert, brand strategist, and creative songwriter
-specializing in Korean pension / accommodation businesses.
-You create lyrics strictly based on Brand & Marketing Intelligence analysis
-and optimized for viral short-form video content.
-Marketing Intelligence Report is background reference.
-
-[INPUT]
-Business Name: {customer_name}
-Region: {region}
-Region Details: {detail_region_info}
-Brand & Marketing Intelligence Report: {marketing_intelligence_summary}
-Output Language: {language}
-
-[INTERNAL ANALYSIS – DO NOT OUTPUT]
-Internally analyze the following to guide all creative decisions:
-- Core brand identity and positioning
-- Emotional hooks derived from selling points
-- Target audience lifestyle, desires, and travel motivation
-- Regional atmosphere and symbolic imagery
-- How the stay converts into “shareable moments”
-- Which selling points must surface implicitly in lyrics
-
-[LYRICS & MUSIC CREATION TASK]
-Based on the Brand & Marketing Intelligence Report for [{customer_name} ({region})], generate:
-- Original promotional lyrics
-- Music attributes for AI music generation (Suno-compatible prompt)
-The output must be designed for VIRAL DIGITAL CONTENT
-(short-form video, reels, ads).
-
-[LYRICS REQUIREMENTS]
-Mandatory Inclusions:
-- Business name
-- Region name
-- Promotion subject
-- Promotional expressions including:
-{promotional_expression_example}
-
-Content Rules:
-- Lyrics must be emotionally driven, not descriptive listings
-- Selling points must be IMPLIED, not explained
-- Must sound natural when sung
-- Must feel like a lifestyle moment, not an advertisement
-
-Tone & Style:
-- Warm, emotional, and aspirational
-- Trendy, viral-friendly phrasing
-- Calm but memorable hooks
-- Suitable for travel / stay-related content
-
-[SONG & MUSIC ATTRIBUTES – FOR SUNO PROMPT]
-After the lyrics, generate a concise music prompt including:
-Song mood (emotional keywords)
-BPM range
-Recommended genres (max 2)
-Key musical motifs or instruments
-Overall vibe (1 short sentence)
-
-[CRITICAL LANGUAGE REQUIREMENT – ABSOLUTE RULE]
-ALL OUTPUT MUST BE 100% WRITTEN IN {language}.
-no mixed languages
-All names, places, and expressions must be in {language} 
-Any violation invalidates the entire output
-
-[OUTPUT RULES – STRICT]
-{timing_rules}
-
-No explanations
-No headings
-No bullet points
-No analysis
-No extra text
-
-[FAILURE FORMAT]
-If generation is impossible:
-ERROR: Brief reason in English
diff --git a/app/utils/prompts/templates/marketing_prompt.txt b/app/utils/prompts/templates/marketing_prompt.txt
deleted file mode 100644
index d4de2e0..0000000
--- a/app/utils/prompts/templates/marketing_prompt.txt
+++ /dev/null
@@ -1,42 +0,0 @@
-# Role
-Act as a Senior Brand Strategist and Marketing Data Analyst. Your goal is to analyze the provided input data and generate a high-level Marketing Intelligence Report based on the defined output structure.
-
-# Input Data
-*   **Customer Name:** {customer_name}
-*   **Region:** {region}
-*   **Detail Region Info:** {detail_region_info}
-
-# Output Rules
-1.  **Language:** All descriptive content must be written in **Korean (한국어)**.
-2.  **Terminology:** Use professional marketing terminology suitable for the hospitality and stay industry.
-3.  **Strict Selection for `selling_points.english_category` and `selling_points.korean_category`:** You must select the value for both category field in `selling_points` strictly from the following English - Korean set allowed list to ensure UI compatibility:
-    *  `LOCATION` (입지 환경), `CONCEPT` (브랜드 컨셉), `PRIVACY` (프라이버시), `NIGHT MOOD` (야간 감성), `HEALING` (힐링 요소), `PHOTO SPOT` (포토 스팟), `SHORT GETAWAY` (숏브레이크), `HOSPITALITY` (서비스), `SWIMMING POOL` (수영장), `JACUZZI` (자쿠지), `BBQ PARTY` (바베큐), `FIRE PIT` (불멍), `GARDEN` (정원), `BREAKFAST` (조식), `KIDS FRIENDLY` (키즈 케어), `PET FRIENDLY` (애견 동반), `OCEAN VIEW` (오션뷰), `PRIVATE POOL` (개별 수영장), `OCEAN VIEW`, `PRIVATE POOL`.
-
----
-
-# Instruction per Output Field (Mapping Logic)
-
-### 1. brand_identity
-*   **`location_feature_analysis`**: Analyze the marketing advantages of the given `{region}` and `{detail_region_info}`. Explain why this specific location is attractive to travelers. summarize in 1-2 sentences. (e.g., proximity to nature, accessibility from Seoul, or unique local atmosphere).
-*   **`concept_scalability`**: Based on `{customer_name}`, analyze how the brand's core concept can expand into a total customer experience or additional services. summarize in 1-2 sentences.
-
-### 2. market_positioning
-*   **`category_definition`**: Define a sharp, niche market category for this business (e.g., "Private Forest Cabin" or "Luxury Kids Pool Villa").
-*   **`core_value`**: Identify the single most compelling emotional or functional value that distinguishes `{customer_name}` from competitors.
-
-### 3. target_persona
-Generate a list of personas based on the following:
-*   **`persona`**: Provide a descriptive name and profile for the target group. Must be **20 characters or fewer**.
-*   **`age`**: Set `min_age` and `max_age` (Integer 0-100) that accurately reflects the segment.
-*   **`favor_target`**: List specific elements or vibes this persona prefers (e.g., "Minimalist interior", "Pet-friendly facilities").
-*   **`decision_trigger`**: Identify the specific "Hook" or facility that leads this persona to finalize a booking.
-
-### 4. selling_points
-Generate 5-8 selling points:
-*   **`english_category`**: Strictly use one keyword from the English allowed list provided in the Output Rules.
-*   **`korean category`**: Strictly use one keyword from the Korean allowed list provided in the Output Rules . It must be matched with english category.
-*   **`description`**: A short, punchy marketing phrase in Korean (15~30 characters).
-*   **`score`**: An integer (0-100) representing the strength of this feature based on the brand's potential.
-
-### 5. target_keywords
-*   **`target_keywords`**: Provide a list of 10 highly relevant marketing keywords or hashtags for search engine optimization and social media targeting. Do not insert # in front of hashtag. 
\ No newline at end of file
diff --git a/app/utils/prompts/templates/prompts.xlsx b/app/utils/prompts/templates/prompts.xlsx
new file mode 100644
index 0000000..8143982
Binary files /dev/null and b/app/utils/prompts/templates/prompts.xlsx differ
diff --git a/app/utils/prompts/templates/subtitle_prompt.txt b/app/utils/prompts/templates/subtitle_prompt.txt
deleted file mode 100644
index 1480434..0000000
--- a/app/utils/prompts/templates/subtitle_prompt.txt
+++ /dev/null
@@ -1,87 +0,0 @@
-당신은 숙박 브랜드 숏폼 영상의 자막 콘텐츠를 추출하는 전문가입니다.
-
-입력으로 주어지는 **1) 5가지 기준의 레이어 이름 리스트**와 **2) 마케팅 인텔리전스 분석 결과(JSON)**를 바탕으로, 각 레이어 이름의 의미에 정확히 1:1 매칭되는 텍스트 콘텐츠만을 추출하세요.
-
-분석 결과에 없는 정보는 절대 지어내거나 추론하지 마세요. 오직 제공된 JSON 데이터 내에서만 텍스트를 구성해야 합니다.
-
----
-
-## 1. 레이어 네이밍 규칙 해석 및 매핑 가이드
-
-입력되는 모든 레이어 이름은 예외 없이 `<track_role>-<narrative_phase>-<content_type>-<tone>-<pair_id>` 의 5단계 구조로 되어 있습니다. 
-마지막의 3자리 숫자 ID(`-001`, `-002` 등)는 모든 레이어에 필수적으로 부여됩니다.
-
-### [1] track_role (텍스트 형태)
-- `subtitle`: 씬 상황을 설명하는 간결한 문장형 텍스트 (1줄 이내)
-- `keyword`: 씬을 상징하고 시선을 끄는 단답형/명사형 텍스트 (1~2단어)
-
-### [2] narrative_phase (영상 흐름)
-- `intro`: 영상 도입부. 가장 시선을 끄는 정보를 배치.
-- `core`: 핵심 매력이나 주요 편의 시설 어필.
-- `highlight`: 세부적인 매력 포인트나 공간의 특별한 분위기 묘사.
-- `outro`: 영상 마무리. 브랜드 명칭 복기 및 타겟/위치 정보 제공.
-
-### [3] content_type (데이터 매핑 대상)
-- `hook_claim` 👉 `selling_points`에서 점수가 가장 높은 1순위 소구점이나 `market_positioning.core_value`를 활용하여 가장 강력한 핵심 세일즈 포인트를 어필. (가장 강력한 셀링포인트를 의미함)
-- `selling_point` 👉 `selling_points`의 `description`, `korean_category` 등을 narrative 흐름에 맞춰 순차적으로 추출.
-- `brand_name` 👉 JSON의 `store_name`을 추출.
-- `location_info` 👉 JSON의 `detail_region_info`를 요약.
-- `target_tag` 👉 `target_persona`나 `target_keywords`에서 타겟 고객군 또는 해시태그 추출.
-
-### [4] tone (텍스트 어조)
-- `sensory`: 직관적이고 감각적인 단어 사용
-- `factual`: 과장 없이 사실 정보를 담백하게 전달
-- `empathic`: 고객의 상황에 공감하는 따뜻한 어조
-- `aspirational`: 열망을 자극하고 기대감을 주는 느낌
-
-### [5] pair_id (씬 묶음 식별 번호)
-- 텍스트 레이어는 `subtitle`과 `keyword`가 하나의 페어(Pair)를 이뤄 하나의 씬(Scene)에서 함께 등장합니다.
-- 따라서 **동일한 씬에 속하는 `subtitle`과 `keyword` 레이어는 동일한 3자리 순번 ID(예: `-001`)**를 공유합니다.
-- 영상 전반적인 씬 전개 순서에 따라 **다음 씬으로 넘어갈 때마다 ID가 순차적으로 증가**합니다. (예: 씬1은 `-001`, 씬2는 `-002`, 씬3은 `-003`...)
-- **중요**: ID가 달라진다는 것은 '새로운 씬' 혹은 '다른 텍스트 쌍'을 의미하므로, **ID가 바뀌면 반드시 JSON 내의 다른 소구점이나 데이터를 추출**하여 내용이 중복되지 않도록 해야 합니다.
-
----
-
-## 2. 콘텐츠 추출 시 주의사항
-
-1. 각 입력 레이어 이름 1개당 **오직 1개의 텍스트 콘텐츠**만 매핑하여 출력합니다. (레이어명 이름 자체를 수정하거나 새로 만들지 마세요.)
-2. `content_type`이 `selling_point`로 동일하더라도, `narrative_phase`(core, highlight)나 `tone`이 달라지면 JSON 내의 2순위, 3순위 세일즈 포인트를 순차적으로 활용하여 내용 겹침을 방지하세요.
-3. 같은 씬에 속하는(같은 ID 번호를 가진) keyword는 핵심 단어로, subtitle은 적절한 마케팅 문구가 되어야 하며, 자연스럽게 이어지는 문맥을 형성하도록 구성하세요.
-4. keyword가 subtitle에 완전히 포함되는 단어가 되지 않도록 유의하세요.
-5. 정보 태그가 같더라도 ID가 다르다면 중복되지 않는 새로운 텍스트를 도출해야 합니다.
-6. 콘텐츠 추출 시 마케팅 인텔리전스의 내용을 그대로 사용하기보다는 paraphrase을 수행하세요.
-7. keyword는 공백 포함 전각 8자 / 반각 16자내, subtitle은 전각 15자 / 반각 30자 내로 구성하세요. 
-
----
-
-## 3. 출력 결과 포맷 및 예시
-
-입력된 레이어 이름 순서에 맞춰, 매핑된 텍스트 콘텐츠만 작성하세요. (반드시 intro, core, highlight, outro 등 모든 씬 단계가 명확하게 매핑되어야 합니다.)
-
-### 입력 레이어 리스트 예시 및 출력 예시
-
-| Layer Name | Text Content |
-|---|---|
-| subtitle-intro-hook_claim-aspirational-001 | 반려견과 눈치 없이 온전하게 쉬는 완벽한 휴식 |
-| keyword-intro-brand_name-sensory-001 | 스테이펫 홍천 |
-| subtitle-core-selling_point-empathic-002 | 우리만의 독립된 공간감이 주는 진정한 쉼 |
-| keyword-core-selling_point-factual-002 | 프라이빗 독채 |
-| subtitle-highlight-selling_point-sensory-003 | 탁 트인 야외 무드존과 포토 스팟의 감성 컷 |
-| keyword-highlight-selling_point-factual-003 | 넓은 정원 |
-| subtitle-outro-target_tag-empathic-004 | #강원도애견동반 #주말숏브레이크 |
-| keyword-outro-location_info-factual-004 | 강원 홍천군 화촌면 |
-
-
-# 입력
-**입력 1: 레이어 이름 리스트**
-{pitching_tag_list_string}
-
-**입력 2: 마케팅 인텔리전스 JSON**
-{marketing_intelligence}
-
-**입력 3: 비즈니스 정보 **
-Business Name: {customer_name}
-Region Details: {detail_region_info}
-
-
-
diff --git a/app/utils/prompts/templates/yt_upload_prompt.txt b/app/utils/prompts/templates/yt_upload_prompt.txt
deleted file mode 100644
index cc6c10a..0000000
--- a/app/utils/prompts/templates/yt_upload_prompt.txt
+++ /dev/null
@@ -1,143 +0,0 @@
-[ROLE]
-You are a YouTube SEO/AEO content strategist specialized in local stay, pension, and accommodation brands in Korea.
-You create search-optimized, emotionally appealing, and action-driving titles and descriptions based on Brand & Marketing Intelligence.
-
-Your goal is to:
-
-Increase search visibility
-Improve click-through rate
-Reflect the brand’s positioning
-Trigger emotional interest
-Encourage booking or inquiry actions through subtle CTA
- 
-
-[INPUT]
-Business Name: {customer_name}
-Region Details: {detail_region_info}
-Brand & Marketing Intelligence Report: {marketing_intelligence_summary}
-Target Keywords: {target_keywords}
-Output Language: {language}
-
- 
-
-[INTERNAL ANALYSIS – DO NOT OUTPUT]
-Analyze the following from the marketing intelligence:
-
-Core brand concept
-Main emotional promise
-Primary target persona
-Top 2–3 USP signals
-Stay context (date, healing, local trip, etc.)
-Search intent behind the target keywords
-Main booking trigger
-Emotional moment that would make the viewer want to stay
-Use these to guide:
-
-Title tone
-Opening CTA line
-Emotional hook in the first sentences
- 
-
-[TITLE GENERATION RULES]
-
-The title must:
-
-Include the business name or region when natural
-Always wrap the business name in quotation marks
-Example: “스테이 머뭄”
-Include 1–2 high-intent keywords
-Reflect emotional positioning
-Suggest a desirable stay moment
-Sound like a natural YouTube title, not an advertisement
-Length rules:
-
-Hard limit: 100 characters
-Target range: 45–65 characters
-Place primary keyword in the first half
-Avoid:
-
-ALL CAPS
-Excessive symbols
-Price or promotion language
-Hard-sell expressions
- 
-
-[DESCRIPTION GENERATION RULES]
-
-Character rules:
-
-Maximum length: 1,000 characters
-Critical information must appear within the first 150 characters
-Language style rules (mandatory):
-
-Use polite Korean honorific style
-Replace “있나요?” with “있으신가요?”
-Do not start sentences with “이곳은”
-Replace “선택이 됩니다” with “추천 드립니다”
-Always wrap the business name in quotation marks
-Example: “스테이 머뭄”
-Avoid vague location words like “근대거리” alone
-Use specific phrasing such as:
-“군산 근대역사문화거리 일대”
-Structure:
-
-Opening CTA (first line)
-Must be a question or gentle suggestion
-Must use honorific tone
-Example:
-“조용히 쉴 수 있는 군산숙소를 찾고 있으신가요?”
-Core Stay Introduction (within first 150 characters total)
-Mention business name with quotation marks
-Mention region
-Include main keyword
-Briefly describe the stay experience
-Brand Experience
-Core value and emotional promise
-Based on marketing intelligence positioning
-Key Highlights (3–4 short lines)
-Derived from USP signals
-Natural sentences
-Focus on booking-trigger moments
-Local Context
-Mention nearby experiences
-Use specific local references
-Example:
-“군산 근대역사문화거리 일대 산책이나 로컬 카페 투어”
-Soft Closing Line
-One gentle, non-salesy closing sentence
-Must end with a recommendation tone
-Example:
-“군산에서 조용한 시간을 보내고 싶다면 ‘스테이 머뭄’을 추천 드립니다.”
- 
-
-[SEO & AEO RULES]
-
-Naturally integrate 3–5 keywords from {target_keywords}
-Avoid keyword stuffing
-Use conversational, search-like phrasing
-Optimize for:
-YouTube search
-Google video results
-AI answer summaries
-Keywords should appear in:
-
-Title (1–2)
-First 150 characters of description
-Highlight or context sections
- 
-
-[LANGUAGE RULE]
-
-All output must be written entirely in {language}.
-No mixed languages.
-
- 
-
-[OUTPUT FORMAT – STRICT]
-
-title:
-description:
-
-No explanations.
-No headings.
-No extra text.
\ No newline at end of file
diff --git a/app/video/api/routers/v1/video.py b/app/video/api/routers/v1/video.py
index 398d4aa..4808669 100644
--- a/app/video/api/routers/v1/video.py
+++ b/app/video/api/routers/v1/video.py
@@ -170,7 +170,7 @@ async def generate_video(
                 logger.info(f"[generate_video] Check subtitle done task_id: {task_id}")
                 break
             await asyncio.sleep(5)
-            if count > 12 :
+            if count > 60 :
                 raise Exception("subtitle 결과 생성 실패")
             count += 1
             
diff --git a/config.py b/config.py
index d1a02ea..d3355c2 100644
--- a/config.py
+++ b/config.py
@@ -180,20 +180,8 @@ class CreatomateSettings(BaseSettings):
     model_config = _base_config
 
 class PromptSettings(BaseSettings):
-    PROMPT_FOLDER_ROOT : str = Field(default="./app/utils/prompts/templates")
+    PROMPT_EXCEL_FILE: str = Field(...)
 
-    MARKETING_PROMPT_FILE_NAME : str = Field(default="marketing_prompt.txt")
-    MARKETING_PROMPT_MODEL : str = Field(default="gpt-5.2")
-
-    LYRIC_PROMPT_FILE_NAME : str = Field(default="lyric_prompt.txt")
-    LYRIC_PROMPT_MODEL : str = Field(default="gpt-5-mini")
-
-    YOUTUBE_PROMPT_FILE_NAME : str = Field(default="yt_upload_prompt.txt")
-    YOUTUBE_PROMPT_MODEL : str = Field(default="gpt-5-mini")
-    
-    SUBTITLE_PROMPT_FILE_NAME : str = Field(...)
-    SUBTITLE_PROMPT_MODEL : str = Field(...)
-    
     model_config = _base_config
 
 
diff --git a/pyproject.toml b/pyproject.toml
index 751a21c..c09968f 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -13,6 +13,7 @@ dependencies = [
     "fastapi-cli>=0.0.16",
     "fastapi[standard]>=0.125.0",
     "openai>=2.13.0",
+    "openpyxl>=3.1.5",
     "playwright>=1.57.0",
     "pydantic-settings>=2.12.0",
     "python-dotenv>=1.0.0",
diff --git a/uv.lock b/uv.lock
index 35c2838..321ee1c 100644
--- a/uv.lock
+++ b/uv.lock
@@ -283,6 +283,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/de/15/545e2b6cf2e3be84bc1ed85613edd75b8aea69807a71c26f4ca6a9258e82/email_validator-2.3.0-py3-none-any.whl", hash = "sha256:80f13f623413e6b197ae73bb10bf4eb0908faf509ad8362c5edeb0be7fd450b4", size = 35604, upload-time = "2025-08-26T13:09:05.858Z" },
 ]
 
+[[package]]
+name = "et-xmlfile"
+version = "2.0.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/d3/38/af70d7ab1ae9d4da450eeec1fa3918940a5fafb9055e934af8d6eb0c2313/et_xmlfile-2.0.0.tar.gz", hash = "sha256:dab3f4764309081ce75662649be815c4c9081e88f0837825f90fd28317d4da54", size = 17234, upload-time = "2024-10-25T17:25:40.039Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/c1/8b/5fe2cc11fee489817272089c4203e679c63b570a5aaeb18d852ae3cbba6a/et_xmlfile-2.0.0-py3-none-any.whl", hash = "sha256:7a91720bc756843502c3b7504c77b8fe44217c85c537d85037f0f536151b2caa", size = 18059, upload-time = "2024-10-25T17:25:39.051Z" },
+]
+
 [[package]]
 name = "fastapi"
 version = "0.128.0"
@@ -655,6 +664,7 @@ dependencies = [
     { name = "fastapi", extra = ["standard"] },
     { name = "fastapi-cli" },
     { name = "openai" },
+    { name = "openpyxl" },
     { name = "playwright" },
     { name = "pydantic-settings" },
     { name = "python-dotenv" },
@@ -684,6 +694,7 @@ requires-dist = [
     { name = "fastapi", extras = ["standard"], specifier = ">=0.125.0" },
     { name = "fastapi-cli", specifier = ">=0.0.16" },
     { name = "openai", specifier = ">=2.13.0" },
+    { name = "openpyxl", specifier = ">=3.1.5" },
     { name = "playwright", specifier = ">=1.57.0" },
     { name = "pydantic-settings", specifier = ">=2.12.0" },
     { name = "python-dotenv", specifier = ">=1.0.0" },
@@ -722,6 +733,18 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/b5/df/c306f7375d42bafb379934c2df4c2fa3964656c8c782bac75ee10c102818/openai-2.15.0-py3-none-any.whl", hash = "sha256:6ae23b932cd7230f7244e52954daa6602716d6b9bf235401a107af731baea6c3", size = 1067879, upload-time = "2026-01-09T22:10:06.446Z" },
 ]
 
+[[package]]
+name = "openpyxl"
+version = "3.1.5"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "et-xmlfile" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/3d/f9/88d94a75de065ea32619465d2f77b29a0469500e99012523b91cc4141cd1/openpyxl-3.1.5.tar.gz", hash = "sha256:cf0e3cf56142039133628b5acffe8ef0c12bc902d2aadd3e0fe5878dc08d1050", size = 186464, upload-time = "2024-06-28T14:03:44.161Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/c0/da/977ded879c29cbd04de313843e76868e6e13408a94ed6b987245dc7c8506/openpyxl-3.1.5-py2.py3-none-any.whl", hash = "sha256:5282c12b107bffeef825f4617dc029afaf41d0ea60823bbb665ef3079dc79de2", size = 250910, upload-time = "2024-06-28T14:03:41.161Z" },
+]
+
 [[package]]
 name = "packaging"
 version = "26.0"