RAG Knowledge Base

The RAG Knowledge Base API is used for document uploading, vector collection maintenance, document retrieval & QA, and session cleanup. Developers only need to use the OPEAI Platform API Key to call the public endpoints.

Authentication method: Authorization: Bearer <API_KEY>

Call Instructions

Base URL: https://api-platform.ope.ai
All RAG endpoints require Bearer Token authentication.

Default Quotas

To ensure the stability of knowledge base construction, retrieval, and QA, the RAG Knowledge Base has the following default quotas:

Limit Item	Default Limit	Description
Knowledge Base Count	Max `5` per account	One API Key can create one independent knowledge base; you can use `5` API Keys to create `5` isolated knowledge bases
Document Count	Max `100` documents per account	Includes uploaded and retained documents in the knowledge base
Total Document Size	Max `200 MB` cumulative size per account	When exceeding the limit, you need to delete some documents, compress/split documents, or apply for a higher quota

For higher capacity, more knowledge bases, enterprise-grade isolation, or dedicated deployment solutions, please contact us.

Recommended Flow

Call Create Knowledge Base Collection to initialize the knowledge base vector collection.
Call Upload Document to write knowledge base documents and complete chunking and indexing.
Call Initiate RAG QA to generate answers based on document content.
Call document query, chunk query, download, delete, and session deletion endpoints as needed.