diff --git a/README.md b/README.md
new file mode 100644
index 0000000..abccfc5
--- /dev/null
+++ b/README.md
@@ -0,0 +1,67 @@
+English | [中文](README_ZH.md)
+
+# Whisper Plugin for Maubot
+
+Whisper Plugin is a plugin for Maubot, Support for transcriptions of audio messages to text in Matrix clients using OpenAI's Whisper API
+
+## Features
+
+- Automatically responds to audio and file messages in rooms
+- Calls OpenAI Whisper API for audio transcription
+- Supports user and room whitelists
+- Support reply in thread
+
+![alt text](sample1.png)
+
+## Installation
+
+1. **Clone or download the plugin code**:
+
+    ```bash
+    git clone <repository_url>
+    cd <repository_directory>
+    zip -9r maubot-stt.mbp *
+    ```
+
+2. **Configure Maubot**:
+
+    Make sure you have installed and set up Maubot. Refer to [Maubot's official documentation](https://docs.mau.fi/maubot/usage/basic.html) for detailed steps.
+
+3. **Upload the plugin**:
+
+    Upload the plugin in the Maubot management interface.
+
+## Configuration
+
+After uploading the plugin, you need to configure it. Here are the configuration items:
+
+- `api_endpoint`: OpenAI Whisper API endpoint, default is `https://api.openai.com/v1/audio/transcriptions`.
+- `openai_api_key`: Your OpenAI API key.
+- `allowed_users`: List of users allowed to use this plugin. If empty, all users are allowed.
+- `allowed_rooms`: List of rooms allowed to use this plugin. If empty, all rooms are allowed.
+
+### Configuration Example
+
+In the Maubot management interface, go to the plugin's configuration page and fill in the following content:
+
+```json
+{
+  "api_endpoint": "https://api.openai.com/v1/audio/transcriptions",
+  "openai_api_key": "your_openai_api_key",
+  "allowed_users": ["@user1:matrix.org", "@user2:matrix.org"],
+  "allowed_rooms": ["!roomid:matrix.org"]
+}
+```
+
+It is recommended to close and reopen the instance configuration after saving.
+
+## Usage
+The plugin will automatically listen to messages in the room and perform the following actions upon receiving a voice message or audio file:
+
+1. Download the audio file.
+2. Call the OpenAI Whisper API for transcription.
+3. Send the transcription result as a text message to the corresponding room.
+
+## License
+
+This project is licensed under the MIT License.
\ No newline at end of file
diff --git a/README_ZH.md b/README_ZH.md
new file mode 100644
index 0000000..61dd519
--- /dev/null
+++ b/README_ZH.md
@@ -0,0 +1,69 @@
+[English](README.md) | 中文
+
+# Whisper Plugin for Maubot
+
+Whisper Plugin 是一个用于 Maubot 的插件，支持在 Matrix 客户端用 OpenAI 的 Whisper API 将音频消息转录为文本。
+
+## 功能
+
+- 自动响应房间中的音频和文件消息
+- 调用 OpenAI Whisper API 进行音频转录
+- 支持用户和房间白名单
+- 支持在线程中回复
+
+![alt text](sample1.png)
+
+## 安装
+
+1. **克隆或下载插件代码**：
+
+    ```bash
+    git clone <repository_url>
+    cd <repository_directory>
+    zip -9r maubot-stt.mbp *
+    ```
+
+2. **配置 Maubot**：
+
+    请确保你已经安装并设置好了 Maubot。具体步骤可以参考 [Maubot 的官方文档](https://docs.mau.fi/maubot/usage/basic.html)。
+
+3. **上传插件**：
+
+    在 Maubot 管理界面上传插件。
+
+## 配置
+
+上传插件后，你需要进行一些配置。以下是配置项的说明：
+
+- `api_endpoint`：OpenAI Whisper API 端点，默认为 `https://api.openai.com/v1/audio/transcriptions`。
+- `openai_api_key`：你的 OpenAI API 密钥。
+- `allowed_users`：允许使用此插件的用户列表。如果为空列表，则允许所有用户使用。
+- `allowed_rooms`：允许使用此插件的房间列表。如果为空列表，则允许所有房间使用。
+
+### 配置示例
+
+在 Maubot 管理界面，进入插件的配置页面，填写以下内容：
+
+```json
+{
+  "api_endpoint": "https://api.openai.com/v1/audio/transcriptions",
+  "openai_api_key": "your_openai_api_key",
+  "allowed_users": ["@user1:matrix.org", "@user2:matrix.org"],
+  "allowed_rooms": ["!roomid:matrix.org"]
+}
+```
+
+建议关闭实例配置保存后再打开。
+
+## 使用
+
+插件会自动监听房间中的消息，当在接收到语音消息或音频文件时进行以下操作：
+
+1. 下载音频文件。
+2. 调用 OpenAI Whisper API 进行转录。
+3. 将转录结果以文本消息的形式发送到相应的房间。
+
+
+## 许可证
+
+本项目采用 MIT 许可证。
\ No newline at end of file
diff --git a/base-config.yaml b/base-config.yaml
new file mode 100644
index 0000000..249b9bb
--- /dev/null
+++ b/base-config.yaml
@@ -0,0 +1,18 @@
+openai_api_key: sk-
+
+# API endpoint to connect to
+whisper_endpoint: https://api.openai.com/v1/audio/transcriptions
+
+# List of allowed users
+# [] means no restriction
+allowed_users: []
+
+# List of allowed rooms
+# [] means no restriction
+allowed_rooms: []
+
+# Optional text to guide the model's style or to continue from a previous audio segment. The prompt should match the language of the audio.
+prompt: "以下是普通话录制的会议记录："
+
+# The language of the input audio. Providing the input language in ISO-639-1 format (Chinese: "zh") will improve accuracy and latency, leave empty by default
+language:
diff --git a/maubot.yaml b/maubot.yaml
new file mode 100644
index 0000000..b8d0706
--- /dev/null
+++ b/maubot.yaml
@@ -0,0 +1,11 @@
+maubot: 0.1.0
+id: nigzu.com.maubot-stt
+version: 0.2.6
+license: MIT
+modules:
+- openai-whisper
+main_class: WhisperPlugin
+config: true
+extra_files:
+- base-config.yaml
+database: false 
diff --git a/openai-whisper.py b/openai-whisper.py
new file mode 100644
index 0000000..bc4aa36
--- /dev/null
+++ b/openai-whisper.py
@@ -0,0 +1,117 @@
+import asyncio
+import json
+import os
+import re
+import aiohttp
+from datetime import datetime
+from typing import Type
+from mautrix.client import Client
+from maubot.handlers import event
+from maubot import Plugin, MessageEvent
+from mautrix.errors import MatrixRequestError
+from mautrix.types import EventType, MessageType, RelationType, TextMessageEventContent, Format,RelatesTo,InReplyTo
+from mautrix.util.config import BaseProxyConfig, ConfigUpdateHelper
+
+class Config(BaseProxyConfig):
+    def do_update(self, helper: ConfigUpdateHelper) -> None:
+        helper.copy("whisper_endpoint")
+        helper.copy("openai_api_key")
+        helper.copy("allowed_users")
+        helper.copy("allowed_rooms")
+        helper.copy("prompt")
+        helper.copy("language")
+
+class WhisperPlugin(Plugin):
+
+    async def start(self) -> None:
+        await super().start()
+        self.config.load_and_update()
+        self.whisper_endpoint = self.config['whisper_endpoint']
+        self.api_key = self.config['openai_api_key']
+        self.prompt = self.config['prompt']
+        self.language = self.config['language']
+        self.allowed_users = self.config['allowed_users']
+        self.allowed_rooms = self.config['allowed_rooms']
+        self.log.debug("Whisper plugin started")
+
+    async def should_respond(self, event: MessageEvent) -> bool:
+        if event.sender == self.client.mxid:
+            return False
+
+        if self.allowed_users and event.sender not in self.allowed_users:
+            return False
+
+        if self.allowed_rooms and event.room_id not in self.allowed_rooms:
+            return False
+
+        return event.content.msgtype == MessageType.AUDIO or event.content.msgtype == MessageType.FILE
+
+    @event.on(EventType.ROOM_MESSAGE)
+    async def on_message(self, event: MessageEvent) -> None:
+        if not await self.should_respond(event):
+            return
+
+        try:
+            await event.mark_read()
+            await self.client.set_typing(event.room_id, timeout=99999)
+
+            audio_bytes = await self.client.download_media(url=event.content.url)
+            transcription = await self.transcribe_audio(audio_bytes)
+
+            await self.client.set_typing(event.room_id, timeout=0)
+
+            content = TextMessageEventContent(
+                msgtype=MessageType.TEXT,
+                body=transcription,
+                format=Format.HTML,
+                formatted_body=transcription
+            )
+            in_reply_to = InReplyTo(event_id=event.event_id)
+            if event.content.relates_to and event.content.relates_to.rel_type == RelationType.THREAD:
+                await event.respond(content, in_thread=True)
+            else:
+                content.relates_to = RelatesTo(
+                    in_reply_to=in_reply_to
+                )
+                await event.respond(content)
+
+        except Exception as e:
+            self.log.exception(f"Something went wrong: {e}")
+            await event.respond(f"Something went wrong: {e}")
+
+
+    async def transcribe_audio(self, audio_bytes: bytes) -> str:
+        headers = {
+            "Authorization": f"Bearer {self.api_key}"
+        }
+        data = aiohttp.FormData()
+        data.add_field('file', audio_bytes, filename='audio.mp3', content_type='audio/mpeg')
+        data.add_field('model', 'whisper-1')
+        if self.prompt:
+            data.add_field('prompt', f"{self.prompt}")
+        if self.language:
+            data.add_field('language', f"{self.language}")
+
+        try:
+            async with aiohttp.ClientSession() as session:
+                async with session.post(self.whisper_endpoint, headers=headers, data=data) as response:
+                    if response.status != 200:
+                        self.log.error(f"Error response from API: {await response.text()}")
+                        return f"Error: {await response.text()}"
+                    response_json = await response.json()
+                    return response_json.get("text", "Sorry, I can't transcribe the audio.")
+        except Exception as e:
+            self.log.exception(f"Failed to transcribe audio, msg: {e}")
+            return "Sorry, an error occurred while transcribing the audio."
+
+    @classmethod
+    def get_config_class(cls) -> Type[BaseProxyConfig]:
+        return Config
+
+    def save_config(self) -> None:
+        self.config.save()
+
+    async def update_config(self, new_config: dict) -> None:
+        self.config.update(new_config)
+        self.save_config()
+        self.log.debug("Configuration updated and saved")
diff --git a/sample1.png b/sample1.png
new file mode 100644
index 0000000..a938125
Binary files /dev/null and b/sample1.png differ