Flowise ​

Flowise is a visual LLM application builder where you create AI workflows by dragging and dropping nodes on a canvas. No coding required — connect LLM nodes, tool nodes, memory nodes, and output nodes with wires to build chatbots, agents, and automation flows. MCP servers appear as tool nodes you can wire into any flow.

Fw
Flowise
Open-Source · flowiseai.com
TRANSPORT
Streamable HTTP ✓
PLATFORM
Web · Self-hosted
MCP VIA
MCP Tool Node

Visual Canvas for MCP-Powered AI Flows ​

Flowise's node-based canvas makes MCP accessible to non-developers. Drag an "MCP Tool" node onto the canvas, paste your server URL, wire it to an agent node, and the agent can call your tools. No Python, no YAML, no JSON config files.

Two flow modes serve different needs. Chatflow mode creates conversational interfaces — chatbots that answer questions using MCP tools for live data. Agentflow mode builds autonomous agents that reason, plan, and execute multi-step tasks using MCP tools at each step.

Highlights:

  • Visual node editor — drag-and-drop AI workflow builder
  • Chatflow mode — conversational interfaces with tool access
  • Agentflow mode — autonomous multi-step agent workflows
  • 90+ integrations — LLMs, vector stores, tools, and memory systems
  • Embeddable chat — deploy flows as chat widgets on any website
  • API endpoints — every flow gets a REST API automatically
  • Self-hosted — Docker or npm, full data control

Drag-and-Drop Setup ​

1. Create a Token ​

In Vinkius Cloud, go to your server → Connection Tokens → Create. Copy the URL.

2. Add MCP Tool Node ​

In the Flowise canvas, drag an MCP Tool node from the toolbox. Paste your Vinkius URL in the configuration panel:

https://edge.vinkius.com/{TOKEN}/mcp

3. Wire to an Agent ​

Connect the MCP Tool node to your Agent or Chatflow node. The agent discovers available tools automatically. Deploy as a chat widget or API endpoint.

FAQ ​

Do I need to code to use MCP in Flowise? No. Everything is visual. Drag the MCP Tool node, paste a URL, wire it to your agent — done.

Can I embed MCP-powered Flowise flows on my website? Yes. Every Flowise flow can be deployed as an embeddable chat widget with a snippet of HTML.

Does Flowise support multiple MCP servers? Yes. Add multiple MCP Tool nodes to the same flow, each connected to a different server. The agent decides which to use.

Is Flowise free? Open-source under Apache 2.0. Self-host at no cost. Flowise Cloud offers managed hosting.