Mastra AI Agent with Apify MCP Server

The Mastra AI Agent with Apify MCP Server is a generic AI agent designed to demonstrate the power of the mastra.ai framework when paired with the Apify MCP Server. This Actor leverages the Model Context Protocol (MCP) to connect AI-driven workflows with Apify’s extensive library of Actors, enabling seamless data extraction, web scraping, and task automation.

Built with TypeScript and the Apify SDK, this Actor serves as a flexible starting point for developers looking to create custom AI agents that tap into real-time web data and tools.

🛠️ Use cases

TikTok trend analysis

Scrape TikTok posts from a specific account and summarize key trends.

Example input:

{
  "prompt": "Scrape the latest 10 posts from @techcrunch on TikTok and summarize key tech topics.",
  "agentName": "TikTok Tech Analyzer",
  "agentInstructions": "You are a tech analyst scraping TikTok posts to identify trends.",
  "modelName": "gpt-4o-mini",
  "actors": ["clockworks/free-tiktok-scraper"]
}

Example output:

{
  "prompt": "Scrape the latest 10 posts from @techcrunch on TikTok and summarize key tech topics.",
  "response": "Here are the latest 10 posts from TechCrunch on TikTok, along with key tech topics summarized..."
}

Web research synthesis

Search the web for a topic and distill top findings.

Example input:

{
  "prompt": "Search the web for recent advances in quantum computing and explain the top 3 breakthroughs.",
  "agentName": "Quantum Research Bot",
  "agentInstructions": "You are a research assistant that scours the web for technical info.",
  "modelName": "gpt-4o",
  "actors": ["apify/rag-web-browser"]
}

Example output:

{
  "prompt": "Search the web for recent advances in quantum computing and explain the top 3 breakthroughs.",
  "response": "Here are the top three recent advances in quantum computing as of 2025..."
}

Your use case 💡

Tweak the input and create your own AI agent to tackle your use case! 🚀

🎯 Features

• 🤖 AI-powered assistance: Uses the mastra.ai framework to process prompts with OpenAI models (e.g., gpt-4o, gpt-4o-mini).
• 🌐 Apify MCP integration: Connects to the Apify MCP Server to access tools like clockworks/free-tiktok-scraper and more.
• 💰 Pay-per-event pricing: Flexible, usage-based billing for Actor runs and task completions.

🔄 How it works

1. 📥 Input

   • Prompt (e.g., "Analyze posts from @openai and summarize AI trends")
   • Agent name (e.g., "Helpful Assistant Agent") - name for the agent
   • Agent instructions (e.g., "You are a helpful assistant...") - description of the agent's role, who the agent is and what it does
   • Model selection (gpt-4o or gpt-4o-mini)
   • Actors (e.g., ["clockworks/free-tiktok-scraper"]) - list of Apify Actors to be served by the MCP server for the agent

2. 🤖 Processing with Mastra

   • Initializes the agent using the mastra.ai framework.
   • Connects to the Apify MCP Server via Server-Sent Events (SSE).
   • Fetches available tools (e.g., TikTok scrapers, web browsers) dynamically.
   • Executes the prompt, leveraging MCP tools as needed.

3. 📤 Output

   • Returns the agent’s response as text.
   • Stores results in an Apify dataset with the prompt and response.

💰 Pricing

This Actor uses the Pay Per Event (PPE) model for cost-effective, usage-based pricing.

Input example

{
  "prompt": "Analyze the posts of @openai and summarize current trends in AI.",
  "agentName": "Social Media Analyst",
  "agentInstructions": "You are a social media analyst who specializes in analyzing posts from various social media platforms.",
  "modelName": "gpt-4o-mini",
  "actors": ["clockworks/free-tiktok-scraper"]
}

Output example

Dataset entry:

{
  "prompt": "Analyze the posts of @openai and summarize current trends in AI.",
  "response": "Here’s a summary of the recent posts from OpenAI's account, highlighting current trends in AI..."
}

🔧 Technical highlights

• Mastra.ai framework: Powers the AI agent.
• Apify MCP Server: Exposes Apify Actors as tools via the Model Context Protocol.
• Apify SDK: Ensures seamless integration with the Apify platform.
• TypeScript: Provides type safety and modern JavaScript features.

🤖 Under the hood with mastra.ai

This Actor uses the mastra.ai framework to create a generic AI agent that interacts with the Apify ecosystem:

🛠️ Tools

Expandable to any Apify Actor from the Apify Store via the MCP Server configuration.

⚙️ Usage

Running on Apify

1. Clone the repository.
2. Push the Actor to the Apify platform:
   hljs language-bash

   Copy

   apify push
   

3. Set Actor OPENAI_API_KEY environment variable.
4. Run the Actor with your input via the Apify Console or API.

Local development

1. Install dependencies:
   hljs language-bash

   Copy

   npm install
   

2. Run locally:
   hljs language-bash

   Copy

   OPENAI_API_KEY=<your-openai-api-key> apify run -i '{"prompt": "Show me latest post from @openai TikTok profile", "actors": ["clockworks/free-tiktok-scraper"]}'
   

📖 Learn more

• Mastra.ai Framework
• Apify MCP Server
• Apify Platform
• Apify SDK Documentation
• What is MCP?
• What are AI agents?
• How to build an AI agent on Apify

🚀 Get started

Ready to build your own AI agent? Deploy this Actor on Apify, tweak the input, and explore the possibilities of mastra.ai and MCP! 🤖✨

🌐 Open source

This Actor is open-source, hosted on GitHub. Want to add features? Open an issue or submit a pull request!