Data Transmission Service:Create and use a DTS RAGFlow knowledge base

Prerequisites

Before you begin, make sure you have:

• A supported vector database instance that meets the requirements below

• An OSS bucket in the same region as the vector database, with Storage Class set to Standard and Storage Redundancy set to Zone-redundant Storage (Recommended)

Supported vector databases

Supported regions

China (Hangzhou), China (Shanghai), China (Beijing), China (Shenzhen), China (Hong Kong), Singapore, and Indonesia (Jakarta).

Usage notes

• Once you enable a public endpoint for a RAGFlow knowledge base, you cannot disable it.

• A RAGFlow account is valid only for the knowledge base it was registered with.

Billing

Billing follows the pay-as-you-go model. For details, see AI Data Preparation metering methods.

Create a RAGFlow knowledge base

1. Go to the RAGFlow knowledge base list page.

   1. Log on to the Data Transmission Service (DTS) console.

   2. In the left navigation pane, click Data preparation.

   3. In the upper-left corner, select the target region.

   4. Click the RAGFlow Knowledge Base tab.

2. Click Create knowledge base.

3. In the Deployment Scope section, enter an Instance Name.

4. In the Network and Zone section, select the VPC, Zone and vSwitch, and Secondary Zone and vSwitch.

   The secondary zone cannot be the same as the primary zone.

5. In the RAGFlow knowledge base configuration section, set the Number of knowledge base services.

   The default Configuration Plan is Default. With the Default plan, you only need to set the number of services — the Knowledge base service specifications are fixed at 4 vCPU, 16 GB Memory.

6. In the Vector database configuration section, set Engine to your vector database type and fill in the required fields.

   Engine: ADB PostgreSQL

   Set Engine to ADB PostgreSQL. Select the target AnalyticDB for PostgreSQL instance as Database. Enter the Database Name, Database schema name, Database account, and Password.

   Engine: PolarSearch

   Set Engine to PolarSearch. Select the PolarDB for MySQL cluster with PolarSearch enabled as Database. Enter the Database account and Password.

   Engine: PolarDB PostgreSQL

   Set Engine to PolarDB PostgreSQL. Select the target PolarDB for PostgreSQL cluster as Database. Enter the Database Name, Database schema name, Database account, and Password.

   Engine: Lindorm

   Set Engine to Lindorm. Select the target Lindorm instance as Database. Enter the Database account and Password.

   Vector database is fixed to Import from existing instance. For ADB PostgreSQL and PolarDB PostgreSQL, Database schema name defaults to public and cannot be changed.

7. In the OSS configuration section, select the OSS Bucket and enter the Path where data will be stored.

8. Click Buy Now.

9. Return to the RAGFlow knowledge base list page and wait until the Running Status changes to Running.

   Click the refresh icon in the upper-right corner to update the list.

Configure an IP whitelist

Before accessing the knowledge base, add the client's IP address to the whitelist.

1. In the Actions column of the target knowledge base, click Configure Whitelist.

2. In the Configure Whitelist panel, add the appropriate IP addresses or CIDR blocks. Separate multiple entries with commas (,). To find your public IP address, run curl ipinfo.io/ip or curl ifconfig.me.

   Access method	IP whitelist to add
   Internal network (client and knowledge base in the same VPC)	Private IP address or CIDR block of the client
   Public network (client on a local server)	Public IP address or CIDR block of the client

3. Click Set.

Log on to RAGFlow

1. In the Actions column of the target knowledge base, click Manage.

   Alternatively, click Log on to knowledge base in the Actions column and select internal or public network access.

2. In the Endpoint section, click Log on with Public Endpoint or Log on with Internal Endpoint.

   Accessing over the public network requires a public endpoint. See Enable a public endpoint.

3. On the RAGFlow login page, enter your email address and password, then click Log on.

4. Manage your knowledge base from the RAGFlow interface.

   For RAGFlow-specific operations, see the official RAGFlow documentation.

   

(Optional) Configure network access for external models

By default, RAGFlow cannot reach external networks. If you want to add a model provider in RAGFlow, configure one of the following options.

Private network: Alibaba Cloud Model Studio

Use PrivateLink to establish a private connection between the VPC and Alibaba Cloud Model Studio. For setup instructions, see Access Alibaba Cloud Model Studio models or application APIs over a private network.

Public network: NAT Gateway

Configure a NAT Gateway for the VPC of the vector database that RAGFlow uses. For setup instructions, see Public NAT Gateway.

Appendix