Xpert Plugin: RAGFlow

Overview

@xpert-ai/plugin-ragflow integrates RAGFlow into the Xpert plugin system, registering a global NestJS module with the host and exposing validation and retrieval capabilities to help you reuse RAGFlow datasets and knowledge bases in your workflows.

Features

• Provides IntegrationRAGFlowPlugin, which logs messages on plugin startup and teardown, and mounts the internal IntegrationRAGFlowModule.
• Offers a RAGFlow integration strategy, publishing name, description, icon, and configuration schema (supports url and apiKey fields) to the Xpert integration directory.
• Built-in RAGFlow knowledge strategy, calling the /api/v1/retrieval endpoint to retrieve specified datasets and returning fragments as LangChain Document + similarity.
• Exposes a POST /ragflow/test validation endpoint for quickly verifying connection info before saving the integration.

Installation

npm install @xpert-ai/plugin-ragflow
# or
pnpm add @xpert-ai/plugin-ragflow

Peer dependencies: The host project must pre-install dependencies such as @xpert-ai/plugin-sdk, @nestjs/common, @metad/contracts, @langchain/core, axios, chalk, lodash-es, and zod. Refer to the repository package.json for specific versions.

Usage

1. Declare the plugin via environment variables when starting the XpertAI service:

   PLUGINS=@xpert-ai/plugin-ragflow

2. When saving the integration, fill in the following configuration fields:

   | Field | Description | | --------- | ------------------------------------------------------------------------------------------- | | url | RAGFlow service base URL, e.g., https://ragflow.your-company.com. The plugin will automatically remove trailing /, /v1, or /api. | | apiKey | API Key for the RAGFlow application, used for request authentication. |

3. When calling the knowledge strategy, the payload should include:

   | Field | Description | | ------------------------- | -------------------------- | | query | User query text. | | k | Number of fragments to return. | | options.knowledgebaseId | RAGFlow dataset ID. |

The plugin returns an array in the form [Document, similarity], which can be directly consumed by downstream LangChain components.

Connection Test Endpoint

Before configuring the integration, you can call the test endpoint to verify if the address and credentials are valid:

curl -X POST https://<your-host>/ragflow/test \
  -H 'Content-Type: application/json' \
  -d '{
     "options": {
        "url": "https://ragflow.your-company.com",
        "apiKey": "ragflow_api_key"
     }
  }'

The service will request the RAGFlow /v1 root path and return the raw response. If parameters are missing or the connection fails, a 400 Bad Request is thrown.

Development & Debugging

Run Nx commands in the monorepo root directory to build and test:

npx nx build @xpert-ai/plugin-ragflow
npx nx test @xpert-ai/plugin-ragflow

nx build compiles TypeScript source code to dist/, and nx test runs unit tests with Jest.

License

This project follows the AGPL-3.0 License in the repository root.