RAG Knowledge Base Management MCP (Management Control Program) Specification Document

1. Service Introduction

This service is a comprehensive RAG (Retrieval-Augmented Generation) knowledge base management MCP (Management Control Program) based on SQLite database, FTS5 full-text search, and sqlite-vec vector search technologies. It provides full lifecycle management functions for knowledge bases including creation, deletion, selection, document import, hybrid search, document management, and database import/export. It supports multiple document formats such as TXT, DOCX, DOC, and PDF, uses asynchronous batch vectorization processing, and combines keyword search and vector search hybrid search strategies to provide a complete solution for knowledge base construction and intelligent retrieval. The service is adapted to both Chinese and English configurations to meet the needs of different scenarios.

2. Core Tool Information

Tool Name	Tool Function	Core Input Parameters	Key Return Content
create_knowledge_base	Creates a new knowledge base. A knowledge base is a container for documents, and each knowledge base can have its own chunk_size and embedding configuration	- kb_name: Knowledge base name (required, must be unique) - chunk_size: Chunk size in tokens (required, e.g., 512, 1024) - embedding_model: Embedding model name (optional) - embedding_endpoint: Embedding service endpoint URL (optional) - embedding_api_key: Embedding service API Key (optional)	Creation result dictionary (including kb_id knowledge base ID, kb_name knowledge base name, chunk_size chunk size)
delete_knowledge_base	Deletes the specified knowledge base. Cannot delete the currently active knowledge base. Deleting a knowledge base will cascade delete all documents and chunks under it	- kb_name: Knowledge base name (required)	Deletion result dictionary (including kb_name deleted knowledge base name)
list_knowledge_bases	Lists all available knowledge bases. Returns detailed information about all knowledge bases, including the currently selected one	No parameters	Knowledge base list dictionary (including knowledge_bases list, count number of knowledge bases, current_kb_id currently selected knowledge base ID. Each knowledge base contains id, name, chunk_size, embedding_model, created_at, is_current, etc.)
select_knowledge_base	Selects a knowledge base as the currently active one. After selection, subsequent operations such as document import and search will be performed in this knowledge base	- kb_name: Knowledge base name (required)	Selection result dictionary (including kb_id knowledge base ID, kb_name knowledge base name, document_count number of documents in this knowledge base)
import_document	Imports documents into the currently selected knowledge base. Supports concurrent import of multiple files. Supports TXT, DOCX, DOC, and PDF formats. Documents will be parsed, split into chunks, and vectors will be generated asynchronously in batch and stored in the database	- file_paths: List of file paths (absolute paths), supports 1~n files (required) - chunk_size: Chunk size in tokens (optional, defaults to knowledge base's chunk_size)	Import result dictionary (including total total number of files, success_count number of successfully imported files, failed_count number of failed files, success_files list of successfully imported files, failed_files list of failed files)
search	Performs hybrid search in the currently selected knowledge base. Combines keyword search (FTS5) and vector search (sqlite-vec), merges results using weighted approach (keyword weight 0.3, vector weight 0.7), deduplicates, reranks using Jaccard similarity, and returns the top-k most relevant results	- query: Query text (required) - top_k: Number of results to return (optional, default from config, usually 5)	Search result dictionary (including chunks chunk list, count number of results. Each chunk contains id, doc_id, content, tokens, chunk_index, doc_name, score, etc.)
list_documents	Lists all documents in the currently selected knowledge base. Returns detailed information about the documents	No parameters	Document list dictionary (including documents document list, count number of documents. Each document contains id, name, file_path, file_type, chunk_size, created_at, updated_at, etc.)
delete_document	Deletes the specified document from the currently selected knowledge base. Deleting a document will cascade delete all chunks of that document	- doc_name: Document name (required)	Deletion result dictionary (including doc_name deleted document name)
update_document	Updates the document's chunk_size and re-parses the document. Will delete existing chunks, re-split the document using the new chunk_size, and asynchronously generate new vectors in batch	- doc_name: Document name (required) - chunk_size: New chunk size in tokens (required)	Update result dictionary (including doc_id document ID, doc_name document name, chunk_count new number of chunks, chunk_size new chunk size)
export_database	Exports the entire kb.db database file to the specified path	- export_path: Export path (absolute path, required)	Export result dictionary (including source_path source database path, export_path export path)
import_database	Imports a .db database file and merges its contents into kb.db. Import will automatically handle name conflicts by adding timestamps to knowledge base and document names	- source_db_path: Source database file path (absolute path, required)	Import result dictionary (including source_path source database path, imported_kb_count number of imported knowledge bases, imported_doc_count number of imported documents)

3. rag-server CLI Usage Guide

rag-server is a command-line wrapper around the core RAG tools, suitable for managing knowledge bases, importing documents, and performing searches directly from the terminal. After installation, you can run rag-server <command> [options] anywhere.

3.1 Basic Usage

• Show help:

rag-server --help

• General format:

rag-server <command> [--option value...]

3.2 Knowledge Base Management Commands

• Create knowledge base

rag-server create_kb --kb_name <name> --chunk_size <size> [--embedding_model model] [--embedding_endpoint URL] [--embedding_api_key KEY]

Description:

• --kb_name: Knowledge base name (required, must be unique)

• --chunk_size: Chunk size in tokens (required, e.g., 512, 1024)

• --embedding_model: Embedding model name (optional)

• --embedding_endpoint: Embedding service endpoint URL (optional)

• --embedding_api_key: Embedding service API Key (optional)

• Delete knowledge base

rag-server delete_kb --kb_name <name>

• List knowledge bases

rag-server list_kb

• Select current knowledge base

rag-server select_kb --kb_name <name>

Note: After selection, the current knowledge base ID is persisted to database/state.json, so different rag-server invocations will share the same selected KB state.

3.3 Document Management Commands

• Import documents

rag-server import_doc --file_paths /abs/path/doc1.txt [/abs/path/doc2.txt ...] [--chunk_size <size>]

Description:

• --file_paths: One or more absolute file paths (required, at least one)

• --chunk_size: Override default chunk size (optional)

• You must call select_kb first to choose an active knowledge base.

• List documents

rag-server list_doc

• Delete document

rag-server delete_doc --doc_name <name>

• Update document chunk size and rebuild vectors

rag-server update_doc --doc_name <name> --chunk_size <new_size>

3.4 Search Commands

• Search

rag-server search --query "<text>" [--top_k <n>]

Description:

• --query: Query text (required)
• --top_k: Number of results to return (optional, default from config, usually 5)

3.5 Database Import/Export Commands

• Export database

rag-server export_db --export_path /abs/path/kb_export.db

• Import database

rag-server import_db --source_db_path /abs/path/other_kb.db

Description: During import, name conflicts of knowledge bases and documents will be automatically resolved by appending timestamps.