DB Chatbot

Rule-based database chatbot for Laravel. It answers shop questions from registered tools and can render PDF and Word reports. There is no LLM and no external AI API.

An LLM driver can be added later by implementing DbChatbot\Chat\Drivers\ChatDriver. The tools stay the same.

Requirements

• PHP 8.3 or newer
• Laravel 13
• A Blade layout component (the default is app-layout)

Install

composer require db-chatbot/laravel

Laravel discovers DbChatbot\DbChatbotServiceProvider.

1. On the host User model, implement DbChatbot\Contracts\ChatUser, use DbChatbot\Concerns\InteractsWithDbChatbot, and cast role to DbChatbot\Enums\Role.
2. Run php artisan migrate. The package adds role on users and the shop, chat, audit, and report tables.
3. Optional: php artisan vendor:publish --tag=db-chatbot-config and change layout when the host layout component is not app-layout.

Authenticated routes:

• GET/POST/DELETE /chat
• GET /admin/audit for admins
• signed GET /reports/{report}

Report files use the reports disk (storage/app/reports). The package registers that disk when the host has not defined it. reports:prune runs daily and deletes files older than 24 hours.

Authorization

Every tool receives the signed-in user through DbChatbot\Contracts\ChatUser and checks a policy or gate itself. Customers can see only their own orders. A missing order and a forbidden order both return Order not found.

Adding a tool

1. Add a class under src/Chat/Tools that implements DbChatbot\Chat\Tools\Tool.
2. Register it in config/chatbot.php. The first matching tool wins.
3. Enforce access inside handle(). Read data with Eloquent and validate every argument.
4. Return a ToolResult. Do not return password hashes, tokens, remember_token, or card data.