Introduction to Mastra AI and Basic Installation Guide

In the booming era of AI development, the demand for open-source platforms that support building machine learning (ML) models is rapidly increasing. Mastra AI emerges as a flexible and easy-to-use tool that helps researchers and AI engineers efficiently build, train, and deploy complex ML pipelines. This article provides an overview of Mastra AI and a basic installation guide to get started.

---

What is Mastra AI?

According to the official documentation (mastra.ai), Mastra is an open-source framework designed to support building, training, and operating AI/ML pipelines at scale.

Mastra is optimized for:

• Managing workflows of complex AI projects.

• Tracking data, models, and experiments.

• Automating the training, evaluation, and deployment processes.

• Supporting customizable and easily extendable plugins.

Mastra aims to become a rapid “launchpad” for AI teams, suitable for both research (R&D) and production-grade systems.

---

Key Components of Mastra

• Pipeline Management: Easily define and manage pipeline steps.

• Experiment Tracking: Record and compare experimental results.

• Deployment Tools: Support for exporting models and deploying them in production environments.

• Plugin System: Integration with external tools like HuggingFace, TensorFlow, and PyTorch.

• UI Dashboard: Visualize processes and results.

---

Basic Installation Guide for Mastra

To install Mastra, you can refer to the detailed guide here:
👉 Mastra Installation Guide

Summary of the basic steps:

---

1. System Requirements

• Node.js v20.0 or higher
• Access to a supported large language model (LLM)

To run Mastra, you need access to an LLM. Typically, you’ll want to get an API key from an LLM provider such as OpenAI , Anthropic , or Google Gemini . You can also run Mastra with a local LLM using Ollama .

---

2.Create a New Project

We recommend starting a new Mastra project using create-mastra, which will scaffold your project. To create a project, run:

npx create-mastra@latest 

On installation, you’ll be guided through the following prompts:

After the prompts, create-mastra will:

1. Set up your project directory with TypeScript
2. Install dependencies
3. Configure your selected components and LLM provider
4. Configure the MCP server in your IDE (if selected) for instant access to docs, examples, and help while you code

MCP Note: If you’re using a different IDE, you can install the MCP server manually by following the instructions in the MCP server docs. Also note that there are additional steps for Cursor and Windsurf to activate the MCP server.

3. Set Up your API Key

Add the API key for your configured LLM provider in your .env file.

OPENAI_API_KEY=<your-openai-key>

Non-Interactive mode:

You can now specify the project name as either a positional argument or with the -p, --project-name option. This works consistently in both the Mastra CLI (mastra create) and create-mastra package. If both are provided, the argument takes precedence over the option.

---

3. Start the Mastra Server

Mastra provides commands to serve your agents via REST endpoints:

mastra run examples/quickstart_pipeline.yaml


Development Server

Run the following command to start the Mastra server:

 npm run dev

If you have the mastra CLI installed, run:

mastra dev

This command creates REST API endpoints for your agents.

---

Test the Endpoint

You can test the agent’s endpoint using curl or fetch:

curl -X POST http://localhost:4111/api/agents/weatherAgent/generate \
-H “Content-Type: application/json” \
-d ‘{“messages”: [“What is the weather in London?”]}’

 

Use Mastra on the Client

To use Mastra in your frontend applications, you can use our type-safe client SDK to interact with your Mastra REST APIs.

See the Mastra Client SDK documentation for detailed usage instructions.

Run from the command line

If you’d like to directly call agents from the command line, you can create a script to get an agent and call it:

Then, run the script to test that everything is set up correctly:

npx tsx src/index.ts

This should output the agent’s response to your console.