RagKit

RagKit is for apps that need knowledge-base capabilities but do not want to build the RAG infrastructure from scratch.

What it solves

If all you have is embeddings, you still need to handle:

• document management
• knowledge-base organization
• synchronization
• query execution
• document lifecycle handling

That usually becomes a lot of engineering work. RagKit provides those services directly.

Best fit

• you need knowledge-base query workflows
• you need document ingestion and management
• you need synchronization and ongoing maintenance of knowledge content
• you do not want to build embeddings, vector-store, and docs lifecycle plumbing yourself

Core capabilities

Base and Document

The core objects in RagKit usually include:

• RagBase
• RagDocument

Document management

Common operations include:

• creating and updating a base
• creating, updating, and deleting documents
• bulk upserting documents

Synchronization

Supported patterns include:

• document-level sync
• base-level sync
• retry after failure

Query

Apps can run queries directly against a specific knowledge base instead of building the retrieval flow from embeddings upward.

When RagKit is a better fit than embeddings alone

• you need a usable knowledge service
• not just vectors that you plan to post-process yourself

Next steps

• Read Choose Your Stack
• Read the RagKit route descriptions in agentos-api README