### Install Project Dependencies

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/事前準備.md

Install all necessary project dependencies from the root directory using npm ci.

```bash
npm ci
```

--------------------------------

### Install Node.js using Mise

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/事前準備.md

Use 'mise' to trust and install the Node.js version specified in the project.

```bash
mise trust
mise install
```

--------------------------------

### Deployment Configuration

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Configuration examples for deploying infrastructure using AWS CDK.

```APIDOC
## AWS CDK Deployment Configuration

### Environment Parameters Example (`self-hosting-prod.ts`)
This file defines the input parameters for the AWS CDK stack for a production self-hosted environment.

```typescript
import { StackInput } from '../lib/stack-input';

export const selfHostingProdParams: Partial<StackInput> = {
  // Basic Settings
  appEnv: 'prod',
  logLevel: 'INFO',

  // Authentication Settings
  selfSignUpEnabled: false,
  samlAuthEnabled: true,
  samlCognitoDomainName: 'genai.auth.ap-northeast-1.amazoncognito.com',
  samlCognitoFederatedIdentityProviderName: 'EntraID',

  // Security Settings
  allowedIpV4AddressRanges: ['203.0.113.0/24', '198.51.100.0/24'],

  // Custom Domain
  hostName: 'genai',
  domainName: 'example.go.jp',
  hostedZoneId: 'Z1234567890ABC',
  certificateArn: 'arn:aws:acm:us-east-1:123456789012:certificate/xxx',

  // Model Settings
  modelRegion: 'ap-northeast-1',
  modelIds: [
    'jp.anthropic.claude-sonnet-4-5-20250929-v1:0',
    'jp.anthropic.claude-haiku-4-5-20251001-v1:0',
    'amazon.nova-pro-v1:0',
    'amazon.nova-lite-v1:0'
  ],
  imageGenerationModelIds: ['amazon.nova-canvas-v1:0'],

  // Monitoring Settings
  monitoring: true,
  slack: {
    enabled: true,
    workspaceId: 'T01234567',
    channelId: 'C01234567'
  }
};
```

### Registering Environment Parameters (`parameter.ts`)
This shows how to register the environment parameters for deployment.

```typescript
import { selfHostingProdParams } from './env-parameters/self-hosting-prod';

const deploy_envs: Record<string, Partial<StackInput>> = {
  '-prod': selfHostingProdParams,
};
```

## Deployment Commands

### Deploy All Stacks
```bash
npm -w packages/cdk run cdk -- deploy --all --require-approval never -c env=-prod
```

### Deploy Specific Stack
```bash
npm -w packages/cdk run cdk -- deploy GenAIWebStack-prod --require-approval never -c env=-prod
```

### Destroy Stacks
```bash
npm -w packages/cdk run cdk destroy -- --all -c env=-prod
```
```

--------------------------------

### Run Local Development Environment

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/ローカル開発環境.md

Execute this command to start the local development server. Replace '環境名' with the appropriate environment name, such as '-dev'.

```shell
sh scripts/run.sh 環境名（例: -dev）
```

--------------------------------

### Sample Request Body Structure

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

An example of how the input fields can be structured for the AI application.

```APIDOC
## Sample Request Body Structure

```json
{
  "question": {
    "title": "Input",
    "desc": "Enter the content you want to ask.",
    "type": "text",
    "required": true,
    "default_value": "Default Value"
  },
  "content": {
    "title": "Content",
    "desc": "Enter content in Markdown format.",
    "type": "textarea"
  },
  "step": {
    "title": "Number Input Example",
    "type": "number",
    "required": true,
    "min": 10,
    "max": 1000,
    "default_value": 20
  },
  "file": {
    "title": "Attached File",
    "type": "file",
    "accept": "image/*"
  },
  "prefecture": {
    "title": "Prefecture",
    "type": "select",
    "items": [
      { "title": "Tokyo", "value": "13" },
      { "title": "Kanagawa", "value": "14" }
    ]
  },
  "fruits": {
    "title": "Fruits",
    "type": "checkbox",
    "items": [
      { "title": "Apple", "value": "apple" },
      { "title": "Banana", "value": "banana" },
      { "title": "Grape", "value": "grape" }
    ]
  },
  "gender": {
    "title": "Gender",
    "type": "radio",
    "items": [
      { "title": "Male", "value": "1" },
      { "title": "Female", "value": "2" }
    ]
  }
}
```
```

--------------------------------

### Asynchronous Processing - Initial Request (With File)

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Example of an initial request for asynchronous processing when file data is included.

```APIDOC
#### Initial Request Example 2: With File

Files are included in `inputs.files` in the format `{"any_key": {"contents": "...", "filename": "..."}}`. `contents` is a Base64 encoded string.

```bash
# Replace with your deployment output
URL="YOUR_INITIAL_REQUEST_URL"
API_KEY="YOUR_API_KEY"

# Base64 encoded file content (e.g., text "hello world")
FILE_CONTENTS=$(echo -n "hello world" | base64)

# Data to send
DATA='''{
  "inputs": {
    "question": "Please summarize the content of this file.",
    "files": [
      { "key": "source_document",
        "contents": "'${FILE_CONTENTS}'",
        "filename": "input.txt"
      }
    ]
  }
}'''

curl -v -X POST \
     -H "Content-Type: application/json" \
     -H "x-api-key: ${API_KEY}" \
     -d "${DATA}" \
     "${URL}"
```

--------------------------------

### Local Development Server Output

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/ローカル開発環境.md

This message indicates the local development server has started successfully. Access http://localhost:5173/ in your browser to verify.

```shell
➜  Local:   http://localhost:5173/
➜  Network: use --host to expose
➜  press h + enter to show help
```

--------------------------------

### Asynchronous Initial Request (With File)

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Example of an initial POST request for asynchronous processing, including a file. The file content must be Base64 encoded and placed within the 'inputs.files' array.

```bash
# デプロイ時の出力に置き換えてください
URL="YOUR_INITIAL_REQUEST_URL"
API_KEY="YOUR_API_KEY"

# Base64エンコードされたファイルコンテンツ (例: "hello world"というテキスト)
FILE_CONTENTS=$(echo -n "hello world" | base64)

# 送信するデータ
DATA='''{
  "inputs": {
    "question": "このファイルの内容を要約してください。",
    "files": [
      { "key": "source_document",
        "contents": "'${FILE_CONTENTS}'",
        "filename": "input.txt"
      }
    ]
  }
}'''

curl -v -X POST \
     -H "Content-Type: application/json" \
     -H "x-api-key: ${API_KEY}" \
     -d "${DATA}" \
     "${URL}"
```

--------------------------------

### Sent Request Example

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Example of the request sent from GenAI Web to the AI application, based on the sample input structure.

```APIDOC
## Sent Request Example

When the sample above is defined, GenAI Web sends the following request to the AI application. Please implement the API to retrieve values under `inputs` for processing.

```json
{
  "inputs": {
    "question": "What is the default value?",
    "content": "Content\nContent",
    "step": 35,
    "files": [
      {
        "key": "key_name",
        "files": [
          {
            "filename": "file_name",
            "content": "base64_data"
          },
          {
            "filename": "file_name",
            "content": "base64_data"
          }
        ]
      },
      {
        "key": "key_name",
        "files": [
          {
            "filename": "file_name",
            "content": "base64_data"
          },
          {
            "filename": "file_name",
            "content": "base64_data"
          }
        ]
      }
    ],
    "prefecture": 13,
    "fruits": "apple,banana",
    "gender": 1
  }
}
```

### Notes

- Numbers are sent as numbers, not strings, even if enclosed in `""`.
- `true` / `false` are sent as booleans, even if enclosed in `""`.
- When multiple checkboxes are selected, they are sent as a comma-separated string like `"a,b,c"`.
  - If a single checkbox is selected, it is sent as a string without a comma, like `"a"`.
- Files are sent as Base64 encoded content.
  - Be aware that validation is performed on the Base64 converted value, even if the file size is less than `max_size`.
```

--------------------------------

### Verify Node.js Version

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/事前準備.md

Check if the installed Node.js version matches the one specified in .node-version or mise.toml.

```bash
node -v
```

--------------------------------

### Asynchronous Initial Request (Text Only)

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Example of an initial POST request for asynchronous processing, sending only text-based inputs. Ensure data is wrapped in an 'inputs' key and includes necessary headers.

```bash
# デプロイ時の出力に置き換えてください
URL="YOUR_INITIAL_REQUEST_URL" # /requests で終わるURL
API_KEY="YOUR_API_KEY"

# 送信するデータ。必ず "inputs" キーでラップすること。
# この形式は後述のカスタマイズポイントで定義します。
DATA='''{
  "inputs": {
    "question": "日本の歴史について、500文字程度で要約してください。",
    "max_length": 512
  }
}'''

curl -v -X POST \
     -H "Content-Type: application/json" \
     -H "x-api-key: ${API_KEY}" \
     -d "${DATA}" \
     "${URL}"
```

--------------------------------

### External AI App Sync Implementation (Server-side)

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Example of how an external API server would implement synchronous AI app processing. It details the expected request format including inputs and optional files.

```bash
# 同期処理の実装例（外部APIサーバー側）
# リクエスト形式
curl -X POST https://your-api.example.com/api/summarize \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "x-user-id: stable-user-id" \
  -d '{
    "inputs": {
      "document": "議事録のテキスト...",
      "summary_level": "standard",
      "files": [
        {
          "key": "attachment",
          "files": [ { "filename": "document.pdf", "content": "base64エンコードされたデータ" } ]
        }
      ]
    },
    "sessionId": "uuid-for-conversation-history"
  }'

# 同期レスポンス
# HTTP 200 OK
# { "outputs": "## 要約結果\n\nMarkdown形式の出力テキスト..." }

```

--------------------------------

### Asynchronous Processing - Initial Request (Text Only)

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Example of an initial request for asynchronous processing when only text data is involved.

```APIDOC
### Asynchronous Processing

#### Initial Request Example 1: Text Only

```bash
# Replace with your deployment output
URL="YOUR_INITIAL_REQUEST_URL" # URL ending with /requests
API_KEY="YOUR_API_KEY"

# Data to send. Must be wrapped with the "inputs" key.
# This format is defined in the customization points described later.
DATA='''{
  "inputs": {
    "question": "Please summarize Japanese history in about 500 characters.",
    "max_length": 512
  }
}'''

curl -v -X POST \
     -H "Content-Type: application/json" \
     -H "x-api-key: ${API_KEY}" \
     -d "${DATA}" \
     "${URL}"
```

--------------------------------

### External AI App Async Implementation (Server-side)

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Example of how an external API server would implement asynchronous AI app processing. It shows the initial accepted response and the structure of status updates.

```bash
# 非同期処理の実装例
# 初期レスポンス (HTTP 202 Accepted)
# {
#   "outputs": "処理を開始しました",
#   "request_id": "req-uuid",
#   "status": "PENDING",
#   "status_url": "/status/req-uuid"
# }

# ステータスエンドポイント (GET /status/{request_id})
# 処理中: { "status": "IN_PROGRESS", "progress": "処理中... 50%" }
# 完了時: {
#   "status": "COMPLETED",
#   "outputs": "処理結果のテキスト",
#   "artifacts": [
#     { "contents": "base64エンコードされたPDF", "display_name": "report.pdf" }
#   ]
# }

```

--------------------------------

### Add System Admin via Script

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/システム管理者設定手順.md

Use this script to add a user to the SystemAdminGroup. Ensure prerequisites like CloudFormation stack deployment and Cognito user registration are met. jq must be installed.

```bash
# parameter.tsのenvキーは `'-dev': devParams,` となっていたらキー値（例では `-dev`）を指定してください
./scripts/add-system-admin.sh <parameter.tsのenvキー> <ユーザー名>

# 例
./scripts/add-system-admin.sh -dev user@example.com

# 成功した場合、以下のような出力が表示されます。
> 完了: ユーザー 'xxxxxxx@example.com' を SystemAdminGroup に追加しました。
```

--------------------------------

### Deploy CDK with Log Sending Enabled

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/ログ設定.md

Deploy the CDK project including the `LoggingStack` which is part of `GenerativeAiUseCasesStack`. Use the `--all` flag to deploy all stacks. Ensure you specify the correct environment, for example, `--c env=-prod` for the production environment.

```bash
npm ci
npm -w packages/cdk run cdk -- deploy --all --require-approval never -c env=-prod
```

--------------------------------

### Get All AI Applications (ExApps) using ExApp API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Retrieve a list of all available AI applications (ExApps) across the platform. This endpoint is useful for discovering and accessing various AI tools.

```typescript
// AIアプリ一覧取得 - GET /exapps
const exAppsResponse = await fetch('/api/exapps', {
  headers: { 'Authorization': `Bearer ${idToken}` }
});
```

--------------------------------

### Start Audio Transcription API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Initiates audio transcription using AWS Transcribe. Supports speaker separation. Requires audio key and optional speaker label and max speakers parameters.

```typescript
// 文字起こし開始 - POST /transcription/start
const startTranscriptionRequest = {
  audioKey: 'path/to/audio.mp3',
  speakerLabel: true,
  maxSpeakers: 3
};

const startResponse = await fetch('/api/transcription/start', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${idToken}`,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify(startTranscriptionRequest)
});
const { jobName } = await startResponse.json();
```

--------------------------------

### Get Team's AI Applications (ExApps) using ExApp API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Fetch a list of AI applications (ExApps) specifically associated with a given team. This allows users to see and manage the AI tools available within their team.

```typescript
// チーム内AIアプリ一覧 - GET /teams/{teamId}/exapps
const teamExAppsResponse = await fetch(`/api/teams/${teamId}/exapps`, {
  headers: { 'Authorization': `Bearer ${idToken}` }
});
```

--------------------------------

### Get Chat - GET /chats/{chatId}

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Fetches details for a specific chat session using its ID. Requires an Authorization header.

```typescript
const getChatResponse = await fetch(`/api/chats/${chatId}`, {
  headers: { 'Authorization': `Bearer ${idToken}` }
});
```

--------------------------------

### ローカル環境からデプロイを実行

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/デプロイ手順.md

ローカル環境からデプロイを実行する際に、`-c env=` パラメータを使用して `deploy_envs` で定義した環境を指定します。これにより、デプロイ先の環境を切り替えることができます。

```bash
npm -w packages/cdk run cdk -- deploy --all --require-approval never -c env=-selfHostingDev
```

--------------------------------

### Status Check (Polling)

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Example of how to poll for the status of an asynchronous request using the 'request_id' obtained from the initial response.

```bash
# 上記で受け取った request_id を使って、処理のステータスを問い合わせます。
# URL と API_KEY は適宜置き換えてください。
REQUEST_ID="a1b2c3d4-e5f6-7890-1234-567890abcdef"
URL="YOUR_STATUS_CHECK_URL"
API_KEY="YOUR_API_KEY"

curl -v -X GET \
     -H "x-api-key: ${API_KEY}" \
     "${URL}/${REQUEST_ID}"
```

--------------------------------

### List Chats - GET /chats

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Retrieves a list of existing chat sessions. Supports pagination using exclusiveStartKey.

```typescript
const listChatsResponse = await fetch('/api/chats?exclusiveStartKey=null', {
  headers: { 'Authorization': `Bearer ${idToken}` }
});
// Response: { "data": [{ "chatId": "...", "title": "...", "updatedDate": "..." }], "lastEvaluatedKey": "..." }
```

--------------------------------

### テンプレートファイルをコピー

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/デプロイ手順.md

環境固有のパラメータ設定ファイルを作成するために、テンプレートファイルをコピーします。`cd` コマンドでディレクトリに移動してから `cp` コマンドを実行してください。

```bash
cd packages/cdk/env-parameters
cp self-hosting-template.ts self-hosting-dev.ts
```

--------------------------------

### List Messages - GET /chats/{chatId}/messages

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Retrieves all messages within a specific chat session using its ID. Requires an Authorization header.

```typescript
const messagesResponse = await fetch(`/api/chats/${chatId}/messages`, {
  headers: { 'Authorization': `Bearer ${idToken}` }
});
// Response: { "messages": [{ "messageId": "...", "role": "...", "content": "...", "createdDate": "..." }] }
```

--------------------------------

### Get AI App History

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Retrieve the execution history for a specific AI application within a team. This endpoint requires the team ID and application ID.

```typescript
// 実行履歴取得 - GET /exapps/history
const historyResponse = await fetch(`/api/teams/${teamId}/exapps/${exAppId}/history`, {
  headers: { 'Authorization': `Bearer ${idToken}` }
});

```

--------------------------------

### Bootstrap AWS CDK

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/事前準備.md

Perform the initial AWS CDK bootstrap for a new AWS account and region combination before deploying.

```bash
npm -w packages/cdk run cdk -- bootstrap
```

--------------------------------

### GitHub Actions 用ワークフローファイルをコピー

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/デプロイ手順.md

GitHub Actions を使用してデプロイする場合、デプロイ用ワークフローファイルのサンプルをコピーして、環境名に合わせてリネームします。その後、ワークフロー内のコメントに従って設定を行ってください。

```bash
cp .github/genai-deploy.yaml.example .github/workflows/genai-deploy-{環境名}.yaml
```

--------------------------------

### Get Signed URL for File Download

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Request a pre-signed URL to download a file from S3 using its key. The file key is typically obtained after a successful upload.

```typescript
// ダウンロード用署名付きURL取得 - GET /file/url
const downloadUrlResponse = await fetch(
  `/api/file/url?bucketName=bucket&filePrefix=${encodeURIComponent(fileKey)}`,
  { headers: { 'Authorization': `Bearer ${idToken}` } }
);
const downloadUrl = await downloadUrlResponse.text();

```

--------------------------------

### Get Signed URL for File Upload

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Request a pre-signed URL from the API to upload a file to S3. The response provides the upload URL and a file key for later reference.

```typescript
// アップロード用署名付きURL取得 - POST /file/url
const uploadUrlResponse = await fetch('/api/file/url', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${idToken}`,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    filename: 'document.pdf',
    mediaFormat: 'application/pdf'
  })
});
const { url: uploadUrl, key: fileKey } = await uploadUrlResponse.json();

```

--------------------------------

### Get Team Members List using Team Management API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Retrieve a list of all users associated with a specific team. This endpoint is useful for auditing and managing team membership.

```typescript
// チームメンバー一覧 - GET /teams/{teamId}/users
const usersResponse = await fetch(`/api/teams/${teamId}/users`, {
  headers: { 'Authorization': `Bearer ${idToken}` }
});
```

--------------------------------

### パラメータ設定ファイルをインポート

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/デプロイ手順.md

デプロイ環境を `parameter.ts` の `deploy_envs` オブジェクトに追加します。これにより、異なる環境へのデプロイが可能になります。

```typescript
import { selfHostingDevParams } from "./env-parameters/self-hosting-dev";

const deploy_envs: Record<string, Partial<StackInput>> = {
  "-selfHostingDev": selfHostingDevParams,
  // 他の環境も追加可能
};

```

--------------------------------

### GET /status/{requestId}

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Retrieves the status of a processing request. The URL should be constructed by appending the request ID to a base URL that ends with '/status/'. An API key is required for authentication.

```APIDOC
## GET /status/{requestId}

### Description
Retrieves the status of a processing request using its unique request ID.

### Method
GET

### Endpoint
`/status/{requestId}`

### Parameters
#### Path Parameters
- **requestId** (string) - Required - The unique identifier for the request.

#### Headers
- **x-api-key** (string) - Required - Your API key for authentication.

### Request Example
```bash
curl -X GET \
     -H "Content-Type: application/json" \
     -H "x-api-key: YOUR_API_KEY" \
     "YOUR_STATUS_CHECK_URL_BASE/status/a1b2c3d4-e5f6-7890-1234-567890abcdef"
```

### Response
#### Success Response (200)
- **request_id** (string) - The unique identifier for the request.
- **status** (string) - The current status of the request (e.g., PENDING, IN_PROGRESS, COMPLETED, ERROR).
- **created_at** (string) - Timestamp when the request was created.
- **updated_at** (string) - Timestamp when the request was last updated.
- **progress** (string) - A human-readable description of the current progress.
- **outputs** (string) - Optional. Text output of the processing, present when status is COMPLETED.
- **artifacts** (array) - Optional. File outputs of the processing, present when status is COMPLETED. Each artifact contains `contents` (Base64 encoded) and `display_name`.
- **error** (object) - Optional. Contains error details (`message`, `details`) when status is ERROR.

#### Response Example (Processing)
```json
{
  "created_at": "2025-08-06T10:00:00.123Z",
  "progress": "処理中... ステップ 3/5",
  "request_id": "a1b2c3d4-e5f6-7890-1234-567890abcdef",
  "status": "IN_PROGRESS",
  "updated_at": "2025-08-06T10:02:30.456Z"
}
```

#### Response Example (Completed)
```json
{
  "artifacts": [
    {
      "contents": "JVBERi0xLjAKMSAwIG9iago8PC9UeXBlL0NhdGFsb2cvUGFnZXMgMiAwIFI+...",
      "display_name": "dummy-report.pdf"
    }
  ],
  "created_at": "2025-08-06T10:00:00.123Z",
  "outputs": "This is a dummy output for question: '日本の歴史について、500文字程度で要約してください。'",
  "progress": "処理が完了しました。結果を保存しています...",
  "request_id": "a1b2c3d4-e5f6-7890-1234-567890abcdef",
  "status": "COMPLETED",
  "updated_at": "2025-08-06T10:05:00.789Z"
}
```

#### Response Example (Error)
```json
{
  "created_at": "2025-08-06T10:00:00.123Z",
  "error": {
    "details": "Required resource not found.",
    "message": "An error occurred during processing."
  },
  "request_id": "a1b2c3d4-e5f6-7890-1234-567890abcdef",
  "status": "ERROR",
  "updated_at": "2025-08-06T10:03:00.500Z"
}
```
```

--------------------------------

### Self-Hosting Production Deployment Parameters

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Configuration for deploying infrastructure using AWS CDK for a self-hosted production environment. Includes settings for app environment, logging, authentication, security, custom domains, AI models, and monitoring.

```typescript
// packages/cdk/env-parameters/self-hosting-prod.ts
import { StackInput } from '../lib/stack-input';

export const selfHostingProdParams: Partial<StackInput> = {
  // 基本設定
  appEnv: 'prod',
  logLevel: 'INFO',

  // 認証設定
  selfSignUpEnabled: false,
  samlAuthEnabled: true,
  samlCognitoDomainName: 'genai.auth.ap-northeast-1.amazoncognito.com',
  samlCognitoFederatedIdentityProviderName: 'EntraID',

  // セキュリティ設定
  allowedIpV4AddressRanges: ['203.0.113.0/24', '198.51.100.0/24'],

  // カスタムドメイン
  hostName: 'genai',
  domainName: 'example.go.jp',
  hostedZoneId: 'Z1234567890ABC',
  certificateArn: 'arn:aws:acm:us-east-1:123456789012:certificate/xxx',

  // モデル設定
  modelRegion: 'ap-northeast-1',
  modelIds: [
    'jp.anthropic.claude-sonnet-4-5-20250929-v1:0',
    'jp.anthropic.claude-haiku-4-5-20251001-v1:0',
    'amazon.nova-pro-v1:0',
    'amazon.nova-lite-v1:0'
  ],
  imageGenerationModelIds: ['amazon.nova-canvas-v1:0'],

  // 監視設定
  monitoring: true,
  slack: {
    enabled: true,
    workspaceId: 'T01234567',
    channelId: 'C01234567'
  }
};

// packages/cdk/parameter.ts への登録
import { selfHostingProdParams } from './env-parameters/self-hosting-prod';
const deploy_envs: Record<string, Partial<StackInput>> = {
  '-prod': selfHostingProdParams,
};

```

--------------------------------

### AWSアクセスキー設定

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリ登録手順書.md

AWS CLI コマンドを実行する前に、`AWS_ACCESS_KEY_ID`、`AWS_SECRET_ACCESS_KEY`、`AWS_SESSION_TOKEN` の環境変数を設定します。これらの認証情報は、API キーを取得するために必要です。

```bash
export AWS_ACCESS_KEY_ID =""
export AWS_SECRET_ACCESS_KEY =""
export AWS_SESSION_TOKEN =""
```

--------------------------------

### System Context API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

API for managing custom system prompts, allowing reuse of AI behavior settings.

```APIDOC
## POST /systemcontexts

### Description
Creates a new system context.

### Method
POST

### Endpoint
/api/systemcontexts

### Request Body
- **systemContext** (object) - Required - The system context object.
  - **systemContextId** (string) - Required - A unique identifier for the system context.
  - **systemContext** (string) - Required - The content of the system prompt.
  - **systemContextTitle** (string) - Required - The title of the system context.

### Request Example
```json
{
  "systemContext": {
    "systemContextId": "some-uuid",
    "systemContext": "You are an AI assistant that helps create administrative documents. Please use polite and accurate Japanese suitable for official documents.",
    "systemContextTitle": "Administrative Document Assistant"
  }
}
```

## GET /systemcontexts

### Description
Retrieves a list of all system contexts.

### Method
GET

### Endpoint
/api/systemcontexts

### Response
#### Success Response (200)
- Returns a list of system context objects.

## PUT /systemcontexts/{systemContextId}/title

### Description
Updates the title of a specific system context.

### Method
PUT

### Endpoint
/api/systemcontexts/{systemContextId}/title

### Parameters
#### Path Parameters
- **systemContextId** (string) - Required - The ID of the system context to update.

### Request Body
- **title** (string) - Required - The new title for the system context.

### Request Example
```json
{
  "title": "New Title"
}
```

## DELETE /systemcontexts/{systemContextId}

### Description
Deletes a specific system context.

### Method
DELETE

### Endpoint
/api/systemcontexts/{systemContextId}

### Parameters
#### Path Parameters
- **systemContextId** (string) - Required - The ID of the system context to delete.
```

--------------------------------

### Create System Context API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Saves and manages custom system prompts for AI behavior settings. Requires system context ID, the prompt itself, and a title.

```typescript
// システムコンテキスト作成 - POST /systemcontexts
const createContextRequest = {
  systemContext: {
    systemContextId: crypto.randomUUID(),
    systemContext: 'あなたは行政文書の作成を支援するAIアシスタントです。公式文書にふさわしい丁寧で正確な日本語を使用してください。',
    systemContextTitle: '行政文書作成アシスタント'
  }
};

await fetch('/api/systemcontexts', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${idToken}`,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify(createContextRequest)
});
```

--------------------------------

### Create AI Application (ExApp) using ExApp API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Register an external AI application (ExApp) within a team. This involves defining its name, endpoint, description, API key, and UI component structure using a JSON schema. The 'placeholder' field dynamically generates forms.

```typescript
// AIアプリ作成 - POST /teams/{teamId}/exapps
const createExAppRequest = {
  exAppName: '議事録要約アプリ',
  endpoint: 'https://api.example.com/summarize',
  description: '会議の議事録を自動的に要約するAIアプリです。',
  howToUse: '1. 議事録テキストを入力\n2. 要約レベルを選択\n3. 実行ボタンをクリック',
  apiKey: 'your-api-key-here',
  copyable: true,
  status: 'published',
  placeholder: JSON.stringify({
    document: {
      type: 'textarea',
      title: '議事録',
      desc: '要約したい議事録のテキストを入力してください',
      required: true,
      min_length: 100,
      max_length: 50000
    },
    summary_level: {
      type: 'select',
      title: '要約レベル',
      items: [
        { title: '簡潔（1段落）', value: 'brief' },
        { title: '標準（3段落）', value: 'standard' },
        { title: '詳細（5段落）', value: 'detailed' }
      ],
      default_value: 'standard'
    },
    include_action_items: {
      type: 'checkbox',
      title: '抽出項目',
      items: [
        { title: 'アクションアイテム', value: 'actions' },
        { title: '決定事項', value: 'decisions' },
        { title: '次回予定', value: 'next_steps' }
      ]
    },
    attachment: {
      type: 'file',
      title: '添付ファイル',
      accept: 'application/pdf,.docx',
      max_size: '10MB',
      multiple: true
    }
  })
};

const exAppResponse = await fetch(`/api/teams/${teamId}/exapps`, {
  method: 'POST',
  headers: { 'Authorization': `Bearer ${idToken}`, 'Content-Type': 'application/json' },
  body: JSON.stringify(createExAppRequest)
});
// Response: { "teamId": "...", "exAppId": "uuid", "exAppName": "議事録要約アプリ", ... }
```

--------------------------------

### Create Chat - POST /chats

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Initiates a new chat session. Requires an Authorization header with an ID token.

```typescript
const createChatResponse = await fetch('/api/chats', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${idToken}`,
    'Content-Type': 'application/json'
  }
});
// Response: { "chat": { "chatId": "uuid", "title": "", "usecase": "chat", "updatedDate": "2024-01-01T00:00:00Z" } }
```

--------------------------------

### Sample API Request Payload

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Example of the JSON payload sent to the AI application based on the defined inputs. Note how different data types are represented, including comma-separated strings for checkboxes and Base64 for files.

```json
{
  "inputs": {
    "question": "デフォルト値とは何ですか?",
    "content": "コンテンツコンテンツ\nコンテンツ",
    "step": 35,
    "files": [
      {
        "key": "key名",
        "files": [
          {
            "filename": "ファイル名",
            "content": "base64データ"
          },
          {
            "filename": "ファイル名",
            "content": "base64データ"
          }
        ]
      },
      {
        "key": "key名",
        "files": [
          {
            "filename": "ファイル名",
            "content": "base64データ"
          },
          {
            "filename": "ファイル名",
            "content": "base64データ"
          }
        ]
      }
    ],
    "prefecture": 13,
    "fruits": "apple,banana",
    "gender": 1
  }
}
```

--------------------------------

### List System Contexts API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Retrieves a list of all saved system contexts. Requires authorization header.

```typescript
// システムコンテキスト一覧取得 - GET /systemcontexts
const contextsResponse = await fetch('/api/systemcontexts', {
  headers: { 'Authorization': `Bearer ${idToken}` }
});
```

--------------------------------

### Get Teams List using Team Management API

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Retrieve a list of teams, optionally filtering by name. This endpoint supports pagination using 'limit' and 'lastEvaluatedKey' parameters. Useful for browsing existing teams.

```typescript
// チーム一覧取得 - GET /teams
const teamsResponse = await fetch('/api/teams?limit=20&name=政策', {
  headers: { 'Authorization': `Bearer ${idToken}` }
});
// Response: { "teams": [...], "lastEvaluatedKey": "..." }
```

--------------------------------

### Add Log Sending Configuration to Parameter File

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/ログ設定.md

Add these settings to your environment's parameter file (e.g., `env-parameters/self-hosting-prod.ts`) to configure the log sending feature. Ensure the `destination.accountId` and `destination.endpointUrl` match your receiving API endpoint details. The `logAccumulationBucketExpirationDays` sets the temporary storage duration, while `DailyExportEventHourUTC` and `TransferS3LogsHourUTC` define the export and transfer times, respectively. The transfer time must be after the export time.

```typescript
export const commonProdParams: Partial<StackInput> = {
  // ... 既存の設定はそのまま ...

  // ▼ ログ送信機能の設定（ここを追加）
  destination: {
    accountId: "123456789012", // 手順 1. でメモした受信側アカウント ID
    endpointUrl: "https://xxx.execute-api.ap-northeast-1.amazonaws.com/prod/", // 手順 1. でメモした API エンドポイント URL
  },
  logAccumulationBucketExpirationDays: 364, // 自アカウント S3 の一時保存期間（日数）
  DailyExportEventHourUTC: "17", // DynamoDB エクスポート時刻（UTC）= JST 02:00
  TransferS3LogsHourUTC: "18", // 受信側への送信時刻（UTC）= JST 03:00
};
```

--------------------------------

### Sample API Input Definition

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/AIアプリAPI仕様.md

Defines various input fields for the AI application, including text, textarea, number, file, select, checkbox, and radio types. Each field has properties like title, description, type, and requirements.

```json
{
  "question": {
    "title": "入力",
    "desc": "質問したい内容を入力してください。",
    "type": "text",
    "required": true,
    "default_value": "デフォルト値"
  },
  "content": {
    "title": "コンテンツ",
    "desc": "マークダウンでコンテンツを入力してください。",
    "type": "textarea"
  },
  "step": {
    "title": "number inputの例",
    "type": "number",
    "required": true,
    "min": 10,
    "max": 1000,
    "default_value": 20
  },
  "file": {
    "title": "添付ファイル",
    "type": "file",
    "accept": "image/*"
  },
  "prefecture": {
    "title": "都道府県",
    "type": "select",
    "items": [
      { "title": "東京都", "value": "13" },
      { "title": "神奈川県", "value": "14" }
    ]
  },
  "fruits": {
    "title": "フルーツ",
    "type": "checkbox",
    "items": [
      { "title": "りんご", "value": "apple" },
      { "title": "ばなな", "value": "banana" },
      { "title": "ぶどう", "value": "grape" }
    ]
  },
  "gender": {
    "title": "性別",
    "type": "radio",
    "items": [
      { "title": "男性", "value": "1" },
      { "title": "女性", "value": "2" }
    ]
  }
}
```

--------------------------------

### Generate Image Variation using Amazon Titan

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Create variations of an existing image using the Amazon Titan Image Generator. Provide the base64 encoded initial image and a text prompt to guide the variation. The output will be a new image based on the input.

```typescript
// 画像バリエーション生成
const variationRequest = {
  model: { type: 'bedrock', modelId: 'amazon.titan-image-generator-v2:0' },
  params: {
    taskType: 'IMAGE_VARIATION',
    textPrompt: [{ text: '同じ構図で秋の風景に変更', weight: 1.0 }],
    initImage: 'base64エンコードされた元画像',
    cfgScale: 7.0,
    height: 1024,
    width: 1024
  }
};
```

--------------------------------

### AWS CDK Deployment Commands

Source: https://context7.com/digital-go-jp/genai-web/llms.txt

Commands for deploying, deploying specific stacks, or destroying infrastructure using AWS CDK. Uses environment parameters for configuration.

```bash
# デプロイコマンド
npm -w packages/cdk run cdk -- deploy --all --require-approval never -c env=-prod

# 特定のスタックのみデプロイ
npm -w packages/cdk run cdk -- deploy GenAIWebStack-prod --require-approval never -c env=-prod

# スタック削除
npm -w packages/cdk run cdk destroy -- --all -c env=-prod
```

--------------------------------

### Create Common App Team Script

Source: https://github.com/digital-go-jp/genai-web/blob/main/docs/共通アプリチームの登録.md

Execute this shell script to create a common app team. Ensure you provide the correct environment key from your parameter.ts file.

```bash
# parameter.tsのenvキーは `'-dev': devParams,` となっていたらキー値（例では `-dev`）を指定してください
./scripts/create-common-app-team.sh <parameter.tsのenvキー>

# 例
# ./scripts/create-common-app-team.sh -dev
```