How to Use the Dify MCP in Mastra AI

Q: How do I register Dify tools within a Mastra AI agent?

You initialize the MCPClient with the server URL, call mcpClient.listTools(), and spread those tools directly into your Mastra agent definition. The framework automatically registers the JSON schemas and makes them executable.

Q: Can I use Mastra AI's human-in-the-loop approval with Dify?

Yes, you can. You set requireToolApproval on the chat or uploadfile tools to pause the Mastra workflow until a human approves the action in your terminal or admin UI.

Q: Where does my application configuration and chat data go?

Your parameter schemas from getparameters and conversation histories from listconversations route through isolated, ephemeral sandboxes. No credentials or prompt configurations are logged or stored on the MCP gateway.

Run complex, multi-step Dify agent workflows with Mastra AI's built-in retry and branching engine.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Dify MCP to Mastra AI

Create your Vinkius account to connect Dify to Mastra AI and route execution through our secure gateway. The platform manages server hosting, runtime updates, and security layers. Configuration requires no manual server provisioning.

GDPR Free for Subscribers

Setup Dify with Mastra AI

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Build Multi-Step Workflows with Mastra AI and Dify

The `chat` tool executes agent runs within Mastra AI's declarative workflow pipelines. You can chain this tool with other services, allowing your agent to process a prompt, evaluate the outcome, and branch to a different path if the response fails validation. Mastra's workflow engine manages the state between these steps automatically. If the Dify endpoint rate-limits your request, Mastra executes an exponential backoff retry. This keeps your autonomous agents running without crashing the entire execution thread. You define the logic once and let the framework handle the operational friction.

Handle Human-in-the-Loop Feedback via MCP Server

The `feedback` tool submits user ratings and corrections back to the Dify control plane directly from your Mastra runtime. This tool bridges the gap between autonomous agent execution and human oversight. You configure Mastra's `requireToolApproval` flag to pause workflows until an operator reviews the agent's output. Once approved, the feedback payload updates the training logs. This loop ensures your models get better over time based on real operational data. You don't need to build a custom dashboard to collect these quality metrics.

Load Dynamic Agent Parameters in Mastra Workflows

The `get_parameters` tool inspects the active Dify application schema before Mastra triggers an agent step. This tool lets your workflow adapt dynamically to changes in the underlying model configuration or system prompts. If a prompt template changes on the platform, your Mastra steps read the new schema instantly. You avoid hardcoding system instructions in your TypeScript files. Instead, you treat the platform as your single source of truth for prompt engineering. This separation of concerns keeps your codebase clean and maintainable.

Setup guide

Set up Dify MCP in Mastra AI

Prerequisites

Node.js 18+ and a TypeScript project
@mastra/mcp + @mastra/core packages
Active Vinkius subscription with a valid endpoint token

1

Install dependencies
Run npm install @mastra/mcp @mastra/core plus your preferred model provider (e.g. @ai-sdk/openai).
2

Configure the MCPClient
Create an MCPClient with your Vinkius endpoint as a URL object. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com.
3

Discover and inject tools
Call mcpClient.listTools() and spread the result into your agent's tools object. All Dify tools become native Mastra tools.
4

Run with any model
Swap openai("gpt-4o") for any AI SDK-compatible provider. Call agent.generate() and the agent routes tool calls through MCP automatically.

agent.ts

import { MCPClient } from "@mastra/mcp";
import { Agent } from "@mastra/core/agent";
import { openai } from "@ai-sdk/openai";

const mcpClient = new MCPClient({
  id: "dify-mcp-client",
  servers: {
    "dify-mcp": {
      url: new URL(
        "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
      ),
    },
  },
});

const agent = new Agent({
  name: "Dify Agent",
  model: openai("gpt-4o"),
  instructions: "You have access to Dify tools.",
  tools: {
    ...(await mcpClient.listTools()),
  },
});

const result = await agent.generate(
  "List recent Dify transactions"
);
console.log(result.text);

Get your connection token →

Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Dify. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.

Why Choose Vinkius

Vinkius connects your tools to AI with real-time monitoring and automatic cost savings — all from one dashboard.

Connect Dify now

Real-time monitoring

Live

visibility into every interaction

Connect your favorite tools to your AI and see exactly what's happening — every request, every response, in real time.

Built-in savings

60%

lower AI costs

Vinkius compresses data between your apps and your AI automatically. Lower bills every month — no configuration required.

Single dashboard

One

place for every integration

Every tool your AI connects to, managed from a single screen. One account, complete control.

Common questions about Dify MCP in Mastra AI

You initialize the `MCPClient` with the server URL, call `mcpClient.listTools()`, and spread those tools directly into your Mastra agent definition. The framework automatically registers the JSON schemas and makes them executable.

Yes, you can. You set `requireToolApproval` on the `chat` or `upload_file` tools to pause the Mastra workflow until a human approves the action in your terminal or admin UI.

Mastra uses built-in retry policies with exponential backoff. If the `chat` tool fails due to a network glitch or rate limit, the workflow retries the call automatically before raising an error.

Yes, it does. The Mastra client automatically detects the transport type from the server URL, supporting both Streamable HTTP and SSE out of the box.

Your parameter schemas from `get_parameters` and conversation histories from `list_conversations` route through isolated, ephemeral sandboxes. No credentials or prompt configurations are logged or stored on the MCP gateway.

Use it with your favorite AI tools

Connect this server to Cursor, Claude, VS Code, and more.

OpenAI Agents SDK sdk-python

Google ADK sdk-python

Pydantic AI sdk-python

Vercel AI SDK sdk-typescript