Before You Start

Overview

Welcome to Pangu Bot, a cutting-edge intelligent dialogue hub designed for advanced large language model (LLM) scenarios. Pangu Bot offers essential conversational capabilities such as intent recognition, error correction, rewriting, and knowledge-enhanced Q&A. It seamlessly integrates with LLMs to deliver a fully functional, end-to-end dialogue platform that is ready to use out of the box.

This document provides detailed descriptions, syntax, parameter descriptions, and examples of Pangu Bot APIs. By using these APIs, you can effortlessly harness the power of Pangu Bot for your applications.

Pangu Bot is currently available only in the CN-Hong Kong region. Pangu Bot is available for trial use. To trial-use it, submit a service ticket.

Endpoints

An endpoint is the request address for calling an API. Endpoints vary depending on services and regions. For more information, see Regions and Endpoints.

Basic Concepts

• Account

  An account is created during user registration and has full access to its associated resources and cloud services. It can reset passwords, assign permissions, and more. The account is the payment entity. For security purposes, do not use it directly for daily management tasks. Instead, create users under the account and delegate routine operations to them.

• User

  A user is created by an account within IAM and represents a person using cloud services. Each user possesses credentials such as passwords and access keys.

  The account name, username, and password are typically required during API authentication processes.

• Region

  A region is a geographic area where cloud services are deployed. Availability zones (AZs) in the same region can communicate with each other over an intranet, while AZs in different regions are isolated from each other. By creating cloud resources in various regions, you can design applications to better meet the needs of specific customers or to comply with legal or other regional requirements.

• AZ

  An AZ consists of one or more physical data centers equipped with independent utility infrastructure. Within an AZ, compute, networking, and storage resources are logically divided into clusters. Multiple AZs within a region are interconnected via high-speed fiber optics, enabling users to build highly available systems across AZs.

• Project

  By default, each region corresponds to a system-predefined project that isolates resources (compute, storage, and networking) between physical regions. Permissions are granted at the project level, allowing users to access all resources within their account for that region. For finer-grained control, sub-projects can be created within the default project. Resources purchased under these sub-projects enable precise permission management, restricting user access to only those resources within the specified sub-project.

  Figure 1 Project isolating model
  

• Checkpoint: A checkpoint marks the latest sequence number of consumed data in an application. When resuming data consumption, this checkpoint ensures continuity from where the process left off.
• App: When multiple applications consume data from the same channel, an app identifier distinguishes each application's checkpoint, ensuring accurate tracking of individual progress.