# Verba MCP for AI Agents AI Agent Connect > Verba connects your RAG platform to your AI agent. It lets you search internal documents, pull back semantic answers with citations, and manage your Weaviate knowledge base directly through your chat interface. No more jumping between tabs to manage your data. ## Overview - **Category:** knowledge-management - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_mosNtker0LQyevIfUh7gGdgJqJOKEbbiUG4YbxIE/ai-agent-connect - **Tags:** rag, semantic-search, document-ingestion, knowledge-base, open-source, retrieval-augmented-generation ## Description This Verba MCP lets you pull your localized knowledge base into your conversational workspace. Instead of manually checking files or navigating a web UI to manage your data, you can just talk to your agent to get things done. You can ask it to find specific information buried in your manuals, and it will pull back summarized answers with exact citations so you know where the info came from. It also handles the heavy lifting of data management. You can add new documents, delete old ones, or check your system health just by typing a command. Because it's part of the Vinkius catalog, you can plug this into your existing workflow without setting up extra bridges. It makes handling complex RAG pipelines feel like a normal conversation rather than a series of API calls. You don't have to worry about the underlying architecture or the complexities of vector embeddings. You just get a reliable way to query your data and keep it updated. Whether you're testing a new set of document chunks or pulling a specific policy from a technical guide, this connection makes your knowledge base accessible in plain English. You can manage your entire library of technical manuals, internal wikis, or product specs without ever touching a terminal or a complex dashboard. This means you can spend your time actually using the information rather than managing the container it sits in. If you need to see if a specific document was successfully ingested, you can ask your agent to list the files or show you the details of a specific entry. It bridges the gap between your data and your daily tasks. It's about making your personal or company knowledge base act as a direct extension of your AI agent's memory, ensuring that the answers you get are grounded in your actual data. ## Tools ### delete_knowledge_document Remove a specific document from your index. Use this to prune dead knowledge permanently. ### get_document_details Show the full content and metadata for a single document. This helps you verify what the agent actually sees. ### list_knowledge_documents List every document your base currently holds. It's the fastest way to get an overview of your indexed files. ### perform_rag_query Execute a RAG query against your knowledge base. It returns summarized answers with citations. ### add_knowledge_document Add new content into your knowledge base. You can include metadata to keep your data organized. ### get_system_config Check your current Verba setup details. This lets you see which embedding models are active and how the system is performing. ## Prompt Examples **Prompt:** ``` Check the status of my local Verba instance and tell me which model it's using. ``` **Response:** ``` I've checked your Verba system configuration. It's currently running on the MiniLM-v12 embedding model. All systems are nominal and ready for queries. ``` **Prompt:** ``` Find the safety protocols in our warehouse manual and tell me the source. ``` **Response:** ``` Based on the 'Warehouse Safety Guide v4', here are the protocols: * Wear steel-toed boots * Use high-vis vests * Keep aisles clear *(Source: DocId: wh-992, confidence: 0.98)* ``` **Prompt:** ``` List all the documents we have in the knowledge base right now. ``` **Response:** ``` You have 24 documents currently indexed in your Verba base. These include the Employee Handbook, Safety Manuals, and the Q3 Project Plans. ``` ## Capabilities ### Retrieve cited answers from your knowledge base Your agent pulls summarized answers with exact document citations from your data. ### Add new documents with specific metadata Insert new content and associated metadata into your knowledge base via chat. ### Delete outdated information from your index Remove specific documents from your knowledge base permanently. ### List every document currently in your base Get a complete list of all documents currently indexed in your Verba system. ### Check the status of your embedding models Verify your current system configuration and embedding model status. ### View the full content of any specific document ID See the full text and metadata for any individual document in your base. ## Use Cases ### Finding specific steps in a massive technical manual An engineer asks for the deployment steps from a 200-page PDF. The agent uses perform_rag_query to pull the exact steps and cites the document ID. ### Cleaning up an old knowledge base A manager realizes some docs are outdated. They ask the agent to list_knowledge_documents and then delete_knowledge_document for the old ones. ### Verifying embedding model status A developer wants to know if the local model is active. They ask the agent to get_system_config to check the current configuration. ### Rapidly updating training data A researcher wants to add new study notes. They tell the agent to add_knowledge_document with specific metadata to keep the RAG updated. ## Benefits - See exactly where your info comes from. Use perform_rag_query to get summarized answers that include specific document citations so you can trust the source. - Manage your data without a web UI. Use add_knowledge_document and delete_knowledge_document to keep your index clean and updated directly from your chat. - Check your system health instantly. Use get_system_config to see if your embedding models and cluster are running correctly without checking logs. - Verify your data contents easily. Use get_document_details to see the raw text and metadata of any specific ID to ensure your index is accurate. - Browse your entire index in one go. Use list_knowledge_documents to keep track of every file your agent can access and stay organized without leaving your chat window. - Speed up your RAG testing cycle. Developers can add text chunks and see retrieval results immediately without leaving their IDE or switching tabs to a management console. ## How It Works The bottom line is you get a chat-based interface for your entire Weaviate RAG stack. 1. Start your local or cloud Verba instance. 2. Plug in your Verba API URL and Key in your client. 3. Ask your agent to query, insert, or list your documents. ## Frequently Asked Questions **Can Verba help me search my own private documents?** Yes, it connects your agent to your private Verba RAG platform. You can search your own files and get answers based only on that data. **Does Verba work with my existing Weaviate database?** Yes, it's designed to work with your Verba instance which runs on Weaviate. It lets your agent interact with that specific knowledge base. **Can I add new info to my knowledge base through chat?** You can use the add_knowledge_document tool. Just tell your agent what to add and it will handle the ingestion for you. **How do I know if the agent is giving me real info?** The perform_rag_query tool provides citations. This means your agent will show you exactly which document it used to generate the answer. **Can I remove old documents from the search results?** Yes, you can use the delete_knowledge_document tool. This allows you to keep your knowledge base clean by removing outdated information. **Is it hard to set up Verba for my AI agent?** It's straightforward. You just need a running Verba instance, an API URL, and an API Key to get started. **Can I query my local Verba instance directly through Cursor?** Yes! Once you configure `VERBA_API_URL` to point to `http://localhost:8000` (or your host port), you can prompt your AI assistant to execute rigorous `perform_rag_query` instructions without ever breaking your developer focus. **How do I insert fresh text data into Verba completely using conversational chat?** Provide the agent with your desired context directly. For example: `Add this chunk of markdown as a new document to Verba: '# Title Content...'`. The agent leverages `addDocumentTool`, serializes the payload, and commits it into Verba's vector store immutably. **Are the query answers backed by citations from its embedded documents?** Absolutely. That's the primary benefit of the integration. When you run `perform_rag_query`, Verba utilizes Weaviate's hybrid search mechanics. The output explicitly includes natural language synthesis backed by the unique document IDs and snippet texts it referenced.