tekBlueprint

tekBlueprint.kbase is the service of a common knowledge base of unifying components from various sources used in tekBlueprint apps.

The main idea of tekBlueprint applications is to minimize duplication and reuse data. For this, “components” are used that can be connected as needed. The service was planned as the main hub for storing and aggregating reusable data. And additionally can be used as an MCP server for integration with LLM.

Key features

• Synchronization of components from different sources – at the moment, synchronization with a customizable list of local catalogs on the file system and with GIT remote repository is supported.
• Integration with tekBlueprint – providing the API to remove components from the knowledge base on request.
• User interface – web interface for searching and viewing the consolidated list of available components with support for full-text search.
• Support for Model Context Protocol (MCP) – for interacting with LLM through the MCP API with the possibility of obtaining information from components from the knowledge base (tools, resources).

Service architecture

The services and components of tekBlueprint.kbase are presented in a simplified form in the diagram below.

Planned services:

• Database –- DBMS for storing metadata of components and repositories.
• Local Cache – local data caching to speed up access to frequently used components.
• FTS – full-text search service for the component database.
• tekBlueprint.kbase – the service itself, providing an API for interacting with external clients.
• IAM – external authentication and authorization service for the minimum necessary user and access management. The service does not yet assume any actions with data and all data is essentially public.
• API:
  • REST API – for interacting with the UI.
  • gRP API – for interacting with console applications.
  • MCP API – for interacting with LLM via the MCP API with the ability to obtain information on components from the knowledge base (tools, resources).
• Data Sources:
  • Web Crawler* [planned] – for collecting information from open sources and updating the knowledge base.
  • tekBlueprint.kbase [planned] – for the ability to build chains of servers.

Access to the knowledge base via MCP for AI agents

tekBlueprint.kbase provides access to the knowledge base via MCP (Model Context Protocol), allowing it to be used by AI agents. This provides:

• Integration with AI agents: The knowledge base is accessible to AI agents via the standard MCP protocol.
• Flexibility and scalability: The knowledge base can be used in various AI applications and systems.
• Simplification of development: Simple agents do not need to develop their own interfaces to access the knowledge base.

Below is an example of using tekBlueprint.kbase with an AI agent.