diff --git a/README.md b/README.md
index 5510b8d..f7867e7 100644
--- a/README.md
+++ b/README.md
@@ -1,6 +1,6 @@
 # NanoComic 漫画分镜生成器
 
-一个面向漫画创作的分镜生成项目：根据用户输入自动调研涉及的人物、事物与场景，并最终输出按分镜逐格的详细漫画脚本（包含镜头/构图、角色动作与表情、穿着、环境与道具、对话及语气等），不输出调研报告。
+一个面向漫画创作的分镜生成项目：根据用户输入自动调研涉及的人物、事物与场景，并最终输出按分镜逐格的详细漫画脚本（包含镜头/构图、角色动作与表情、穿着、环境与道具、对话及语气等）。
 
 ## 致谢
 
diff --git a/backend/src/agent/prompts.py b/backend/src/agent/prompts.py
index 9a6c1ea..e17ecf4 100644
--- a/backend/src/agent/prompts.py
+++ b/backend/src/agent/prompts.py
@@ -86,15 +86,15 @@ Strict Requirements:
 - Output ONLY valid JSON. No prose, no markdown fences, no comments.
 - The JSON must be an object where EACH KEY (string) represents a page (e.g., "page_1", "page_2").
 - Each page value must be an object with EXACTLY two keys:
-  - "id": string, the page identifier (must equal the key)
+  - "id": integer, the page identifier (use 1-based numbers)
   - "detail": string, a thorough page description that fine-grains every panel: characters' actions, attire, environment, camera/framing, dialogue with tone, props, transitions.
 - Do NOT invent facts. Ground all details in the provided summaries.
 - Include sources inline inside the "detail" string using markdown links (e.g., [apnews](https://vertexaisearch.cloud.google.com/id/1-0)).
 
 Example JSON (structure only):
 {
-  "page_1": { "id": "page_1", "detail": "..." },
-  "page_2": { "id": "page_2", "detail": "..." }
+  "page_1": { "id": 1, "detail": "..." },
+  "page_2": { "id": 2, "detail": "..." }
 }
 
 Instructions: