# VectorShift MCP > VectorShift provides full control over complex AI automation and Retrieval-Augmented Generation (RAG) workflows. Use this MCP to manage entire data pipelines, query internal knowledge bases with semantic search, or deploy and interact with custom chatbots—all directly from your agent's conversation. ## Overview - **Category:** knowledge-management - **Price:** Free - **Tags:** rag, ai-workflows, automation, knowledge-base, llm-ops ## Description Need to run complicated LLM processes? This connector lets you treat AI automation like any other service: manage it via natural language commands. You can build complex, multi-step workflows that execute data transformations before querying a knowledge base or starting a chatbot session. For example, your agent can first query the available chatbots using `list_chatbots`, then upload context files with `upload_chatbot_files` and run the conversation with `run_chatbot`. If you need to make sure those processes are reliable across different services—say, linking this automation layer to a billing MCP or a messaging service—Vinkius lets your agent chain them together using one connection. This means you build massive automations without worrying about platform boundaries. You're running everything from the initial setup (`create_pipeline`) through monitoring (like pausing a workflow with `pause_pipeline` or terminating it via `terminate_pipeline`). ## Tools ### bulk_run_pipeline Runs multiple instances of a defined workflow simultaneously for high-volume processing. ### create_chatbot Initializes and provisions a new, dedicated chatbot instance. ### create_knowledge_base Sets up the container for storing indexed organizational knowledge. ### create_pipeline Defines and builds a new multi-stage, automated data workflow. ### create_transformation Builds reusable logic blocks to clean or reshape structured data (Python/JS). ### delete_chatbot Removes an existing chatbot instance entirely from the system. ### delete_knowledge_base_documents Removes specific documents from a knowledge base using their unique IDs. ### delete_knowledge_base Permanently removes a knowledge base container and its associated data. ### delete_pipeline Deletes an entire workflow pipeline definition by its ID. ### delete_transformation Removes a custom data transformation logic block. ### get_chatbot Retrieves the details of a chatbot using either its ID or name. ### get_knowledge_base Fetches the metadata and status of a knowledge base by ID or name. ### get_pipeline Retrieves the full definition and configuration of a specific pipeline workflow. ### get_transformation Gets the current details and code for a defined data transformation logic. ### index_knowledge_base Adds files, URLs, or documents to be processed and stored within a knowledge base. ### list_chatbots Returns a list of all chatbot instances currently available for use. ### list_knowledge_base_documents Finds and lists the specific documents stored within a knowledge base container. ### list_knowledge_bases Returns a list of all configured knowledge bases in your account. ### list_pipelines Lists every defined workflow pipeline that you have set up. ### list_transformations Retrieves a list of all custom data transformation scripts available. ### pause_pipeline Stops a currently executing pipeline workflow instance temporarily. ### query_knowledge_base Performs a semantic search against the knowledge base to find relevant context for your agent. ### resume_pipeline Restarts one or more pipeline instances that were previously paused. ### run_chatbot Sends a specific message to a chatbot and waits for the generated response. ### run_pipeline Executes an entire pipeline workflow with specified inputs, starting its process. ### run_transformation Runs a saved data transformation script using specific input variables. ### terminate_chatbot Immediately ends an active chatbot session that is currently running. ### terminate_pipeline Abruptly stops a running pipeline workflow instance when it's no longer needed. ### upload_chatbot_files Upload files to a chatbot session ## Prompt Examples **Prompt:** ``` List all my available VectorShift pipelines. ``` **Response:** ``` I've retrieved your pipelines. You have 3 active workflows: 'Customer Support Bot' (ID: pipe_1), 'Data Extractor' (ID: pipe_2), and 'Lead Scraper' (ID: pipe_3). ``` **Prompt:** ``` Search the 'Company Wiki' knowledge base (ID: kb_99) for 'remote work policy'. ``` **Response:** ``` Searching... I found relevant sections: 'Employees can work remotely up to 3 days a week' and 'Home office stipends are processed monthly'. Would you like more details? ``` **Prompt:** ``` Run the 'Data Extractor' pipeline (ID: pipe_2) with the input 'url: https://example.com'. ``` **Response:** ``` Pipeline 'Data Extractor' started. The execution is in progress. I will notify you once the data extraction from example.com is complete. ``` ## Capabilities ### Manage Workflow Pipelines Create, run, and control complex, multi-step data pipelines that execute custom logic. ### Build Knowledge Bases Index documents from files or URLs to create searchable knowledge bases for grounded AI responses. ### Operate Chatbots Deploy, manage, and send messages to specialized chatbots directly through your agent. ### Execute Data Logic Run custom data transformations using Python or JavaScript logic as part of a larger process. ### Monitor Execution State List and control running workflows, allowing you to pause, resume, or terminate instances. ## Use Cases ### The Product Team Needs Technical Answers A product manager needs to know the specific rules for a new feature. Instead of digging through shared drives, they ask their agent to `query_knowledge_base` using the 'Product Wiki' knowledge base. The agent returns precise text snippets found via semantic search. ### The Data Team Needs Batch Processing A data analyst has 50 spreadsheets that need standardization before being loaded into a new system. They define a transformation with `create_transformation` and then use the agent to execute it via `run_transformation` on all files, handling the batch process. ### The Support Team Needs an Internal Bot Support needs instant access to HR policies. They first create a knowledge base with HR docs using `index_knowledge_base`, then use the agent to deploy and test it by calling `create_chatbot` and running the conversation through `run_chatbot`. ### The DevOps Team Needs Full Visibility A developer needs to test a complex data flow. They first define a pipeline using `create_pipeline`, then run it with initial inputs via `run_pipeline`. If the job fails, they can immediately check if the instance is running using `get_pipeline`. ## Benefits - Need to ground an agent in proprietary information? You can `index_knowledge_base` using files or URLs, and then use that context immediately when running a query with `query_knowledge_base`. - Stop relying on manual data prep. By defining custom logic using `create_transformation`, you ensure the input for any workflow is always clean before execution. - Don't just run a chatbot; manage its entire lifecycle. Use `list_chatbots` to see what’s available and `get_chatbot` to confirm its setup. - Handle large-scale jobs without constant babysitting. You can define a pipeline with `create_pipeline` and then use `bulk_run_pipeline` to fire off dozens of tasks at once. - When things go sideways, you have control. If a workflow is taking too long, you can `pause_pipeline`, inspect it, and restart later using `resume_pipeline`. ## How It Works The bottom line is that it gives your AI client the full operational controls—the equivalent of an admin dashboard—to manage all its data sources and workflows. 1. Subscribe to the MCP and provide your VectorShift API key. 2. Your agent uses commands (like 'list all pipelines') to see what resources you have created. 3. You tell your agent which action to take, like running a query against a specific knowledge base ID. ## Frequently Asked Questions **How do I manage multiple workflows with VectorShift (AI Workflow & RAG Automation)?** You use the `list_pipelines` tool to see every workflow you've defined. You can then choose which one to run or modify using tools like `get_pipeline`. **Is VectorShift (AI Workflow & RAG Automation) good for batch processing?** Yes, it handles volume well. Instead of running pipelines one by one, you can use the `bulk_run_pipeline` tool to execute multiple instances in parallel. **What's the difference between a chatbot and a knowledge base?** A knowledge base (`create_knowledge_base`) is just the data repository. A chatbot requires you to create it using `create_chatbot`, which allows for active conversation management via tools like `run_chatbot`. **How do I stop a pipeline that's running too long?** You use the `terminate_pipeline` tool. This immediately stops any workflow instance, preventing unnecessary resource consumption. **What data types can I use with the `index_knowledge_base` tool in VectorShift?** It accepts multiple formats, including raw files and URLs. You simply point it at the source content, and VectorShift handles turning that material into searchable vectors for your knowledge base. **How secure is my data when I use VectorShift (AI Workflow & RAG Automation) with my AI client?** Security relies on a zero-trust proxy for credentials. Your keys are never saved to disk, and every single tool call generates a cryptographically signed audit trail so you can trace exactly what happened. **What is the purpose of using the `create_transformation` tool in VectorShift?** This tool lets you write custom logic using Python or JavaScript. You use it to clean, format, or manipulate data inputs, ensuring they are perfectly structured before a pipeline consumes them. **If I run a workflow with `run_pipeline`, can I monitor or pause its progress?** Absolutely. The system tracks running pipelines for you. You have tools available to temporarily halt execution, like `pause_pipeline`, and then resume the process later without losing your state. **How do I search for specific information within my VectorShift knowledge base?** Use the `query_knowledge_base` tool with your Knowledge Base ID and the search query. The agent will perform a semantic search and return the most relevant data chunks. **Can I trigger a specific AI workflow with custom parameters?** Yes! Use the `run_pipeline` tool. Provide the Pipeline ID and a JSON object mapping your input names to their respective values to start the execution. **Is it possible to add new documents to a knowledge base through the agent?** Absolutely. Use the `index_knowledge_base` tool to add data (such as URLs or file content) to an existing knowledge base for real-time RAG updates.