# FlowiseAI MCP MCP

> FlowiseAI connects your agent to a self-hosted LLM platform, giving you direct control over complex AI workflows and RAG pipelines. You can trigger specific chatflows with natural language, manage vector data ingestion, and monitor performance metrics—all through simple conversation. This MCP turns the back end of your chatbot into an actionable tool for your agent.

## Overview
- **Category:** friends-mcp
- **Price:** Free
- **Tags:** llm-workflows, rag-pipelines, chatbot-development, visual-programming, ai-agents, low-code

## Description

Building sophisticated AI chatbots usually means juggling a bunch of UIs: one for testing flows, another for managing documents, and yet another for checking credentials. With this MCP, you skip all that clicking. Your agent talks directly to your Flowise setup, letting it run complex logic and interact with your LLM orchestration layer entirely through chat commands. You can ask the system to execute a specific prediction flow using simple natural language. Need to add new knowledge? Just tell your agent to push data into the vector store. Want to see what credentials you've set up or list all existing workflows? It pulls that structure for you. This means your AI client becomes more than just a chatbot; it’s an operational coordinator, giving you total visibility over everything running in your Flowise ecosystem, right within your preferred Vinkius-connected environment.

## Tools

### get_chatflow_details
Retrieves detailed technical information about one chatflow.

### get_server_version
Returns the current operational version of the Flowise server.

### list_ai_assistants
Lists available OpenAI-style assistants configured in the system.

### list_chatflows
Lists all existing LLM orchestration flows accessible to the agent.

### list_flowise_credentials
Lists the configured API credentials for the Flowise environment.

### list_chat_feedback
Retrieves a list of user feedback submitted for a specific chatflow.

### list_flow_leads
Lists records of captured leads generated by the chatflows.

### list_marketplace_templates
Retrieves a list of available chatflow templates from the marketplace.

### list_external_tools
Retrieves a list of custom tools defined for the system.

### list_flow_variables
Retrieves a list of global variables used across all flows.

### execute_chatflow_prediction
Runs a specific LLM flow prediction based on natural language input.

### upsert_vector_data
Pushes structured or raw data into the associated vector store.

## Prompt Examples

**Prompt:** 
```
List all my chatflows in Flowise.
```

**Response:** 
```
I've retrieved your LLM flows. You have 3 active chatflows: 'Customer Support Bot' (ID: cf_1), 'RAG Assistant' (ID: cf_2), and 'Lead Generator' (ID: cf_3). Which one would you like to inspect?
```

**Prompt:** 
```
Execute chatflow 'cf_1' with question: 'How do I reset my password?'
```

**Response:** 
```
Prediction complete! The bot responds: 'To reset your password, please go to the login screen and click on "Forgot Password". Check your email for further instructions.' Would you like to review the session memory?
```

**Prompt:** 
```
Upsert this data into vector store for chatflow 'cf_2': [data]
```

**Response:** 
```
Data ingested! I've successfully pushed the documents to the vector store associated with chatflow 'cf_2'. The new context is now available for RAG-based predictions.
```

## Capabilities

### Run Chatflow Predictions
Sends natural language input to a specific LLM flow and returns the generated response.

### List Available Workflows
Retrieves a list of all structured AI orchestration flows you've built.

### Inspect Flow Details
Pulls the complete technical structure and metadata for any specific chatflow.

### Ingest Knowledge Data
Programmatically pushes raw documents or data into your linked vector store for RAG context.

### View System Assets
Lists configured credentials, global variables, and custom tools used by the platform.

### Monitor Operational Metrics
Retrieves information like captured leads, user feedback, and active assistant profiles.

## Use Cases

### Debugging an AI Agent
An AI developer needs to know why a complex bot failed. Instead of clicking through the UI, they ask their agent to run `list_chatflows` to get the IDs, then use `get_chatflow_details` on the suspected flow ID to see exactly where the logic might be broken.

### Updating Knowledge Bases
A data team has just processed a new batch of legal documents. They instruct their agent to run `upsert_vector_data` immediately, pushing the raw content into the vector store associated with the 'Compliance Bot' chatflow.

### Reviewing Marketing Funnels
A product manager needs a performance report. They ask their agent to use `list_flow_leads`, which instantly returns all captured leads and shows how many users interacted with the 'Lead Generator' flow.

### Auditing System Settings
An operations engineer wants to verify what credentials are active. They simply ask their agent to use `list_flowise_credentials`, getting a clear, conversational list of all connected services.

## Benefits

- You don't have to manually check credentials. Use the `list_flowise_credentials` tool to instantly review all configured API keys and service tokens through a simple query.
- Monitoring is fast. Instead of jumping into separate dashboards, you can use your agent to list captured leads with `list_flow_leads` and gather user feedback via `list_chat_feedback` in one conversation.
- Testing flows becomes conversational. You bypass the Flowise UI by using `execute_chatflow_prediction`, letting your AI client run a flow and report the result immediately.
- Data ingestion is automated. Use `upsert_vector_data` to feed new documents into the vector store, ensuring your RAG pipelines have current knowledge without leaving your chat interface.
- System oversight is simple. The `list_chatflows` tool gives you a complete inventory of all active workflows, while `get_chatflow_details` lets you inspect their full technical makeup.

## How It Works

The bottom line is you manage your entire LLM data pipeline and flow logic using conversational commands, without ever leaving your primary AI client interface.

1. Subscribe to this MCP and provide your Flowise Instance URL and API Key.
2. Your agent authenticates the connection, giving it immediate programmatic access to all defined endpoints.
3. You issue a natural language command—like 'list all chatflows' or 'run prediction for X'—and the agent executes the required tool call.

## Frequently Asked Questions

**How do I check credentials using the FlowiseAI MCP?**
You use the `list_flowise_credentials` tool. This instantly retrieves a list of all configured API keys and service tokens, letting you verify your connections without navigating the platform's settings.

**Is `upsert_vector_data` for updating documents?**
Yes, it is designed specifically to push new data or updated documents into the vector store. This action ensures that the Retrieval-Augmented Generation (RAG) context is always current.

**What if I need to run a prediction on an old flow?**
First, use `list_chatflows` to confirm the flow ID. Then, pass that ID and your prompt to `execute_chatflow_prediction`. Your agent handles the execution.

**How do I list all my flows using the FlowiseAI MCP?**
Run the `list_chatflows` tool. It provides a clean, conversational summary of every active LLM orchestration flow you have set up in your environment.

**How do I monitor user feedback for a chatflow using list_chat_feedback?**
You can retrieve specific user comments and performance data by calling list_chat_feedback. This lets you see how users are interacting with the flow, pinpointing confusing spots or common queries that need model refinement.

**What is the purpose of get_server_version in this MCP?**
This tool reports the specific version of your Flowise instance. Checking the server version ensures your AI client and the underlying infrastructure are compatible, helping you quickly diagnose setup discrepancies.

**Where can I check the available context parameters using list_flow_variables?**
list_flow_variables displays all global variables configured for your entire ecosystem. This is essential for understanding what external data or custom settings are accessible to every single chatflow running through the MCP.

**How do I view the internal components of a flow using get_chatflow_details?**
get_chatflow_details pulls up the complete technical structure for one specific chatflow. This allows developers to inspect every connected component, variable, and logic branch without needing access to the original Flowise UI.

**How do I find my API Key in Flowise?**
Log in to your Flowise dashboard and click on the **API Keys** tab in the sidebar to generate or copy your unique token.

**Does this support multi-tenant instances?**
Yes! Ensure you provide the full Instance URL and the API Key corresponding to the specific environment you want to manage.

**Can I push documents to vector stores via AI?**
Absolutely. Use the `upsert_vector_data` tool by providing the `chatflow_id` and the JSON payload containing your document data.