Getting Started with AgentStack

Welcome to AgentStack (formerly Mastra)! This comprehensive multi-agent framework helps you build production-grade AI applications with 22+ specialized agents, 30+ enterprise tools, 10+ workflows, and 4 agent networks.

Installation

Install dependencies for the full AgentStack framework:

npm install

Quick Start

Start the development server with Mastra CLI:

npm run dev

This starts both the Next.js frontend and Mastra backend services concurrently.

Core Architecture

AgentStack is built around several key components:

🤖 Agents (22+ Available)

Research Agent - Web research and data gathering
Content Strategist - Content planning and optimization
Copywriter Agent - Professional content creation
Data Processing Agents - CSV, JSON, and document processing
Financial Agents - Stock analysis and reporting
Script Writers - Automated content generation
Calendar & Weather Agents - External API integrations

🛠️ Tools (30+ Enterprise Tools)

Financial Tools: Polygon, Finnhub, Alpha Vantage APIs
Research Tools: SerpAPI, ArXiv integration
Data Tools: CSV/JSON processing, document chunking
RAG Tools: Vector embeddings, semantic search
Development Tools: GitHub integration, code analysis

🔄 Workflows (10+ Pre-built)

Content Review Workflow - Multi-stage content validation
Financial Report Workflow - Automated financial analysis
Research Synthesis - Knowledge aggregation
Document Processing - File analysis and extraction
Stock Analysis - Market intelligence
Learning Extraction - Educational content processing

🌐 Networks (4 Agent Networks)

Agent Network - Primary router: routes to research, stock, weather, content agents
Data Pipeline Network - ETL operations: CSV/JSON import/export/transform
Report Generation Network - Multi-source reporting: research → transform → report
Research Pipeline Network - Research lifecycle: arXiv → PDF → chunk → index → query

Environment Setup

Configure your environment variables in .env:

# Database
SUPABASE=postgresql://your-connection-string

# AI Providers
OPENAI_API_KEY=sk-your-key
GOOGLE_API_KEY=your-key
ANTHROPIC_API_KEY=your-key

# Tools
SERPAPI_API_KEY=your-key
ALPHA_VANTAGE_API_KEY=your-key
FINNHUB_API_KEY=your-key

Basic Usage

Creating an Agent

import { Agent } from '@mastra/core/agent'
import { googleAI3 } from '@mastra/config/google'
import { Memory } from '@mastra/memory'

const myAgent = new Agent({
    id: 'my-agent',
    name: 'My Custom Agent',
    description: 'A custom agent for specific tasks',
    instructions: 'You are a helpful assistant...',
    model: googleAI3,
    memory: pgMemory,
    tools: {
        // Add your tools here
    },
})

const result = await myAgent.stream('Hello world')

Creating a Workflow

import { createWorkflow, createStep } from '@mastra/core/workflows'
import { z } from 'zod'

const myStep = createStep({
    id: 'my-step',
    inputSchema: z.object({ input: z.string() }),
    outputSchema: z.object({ output: z.string() }),
    execute: async ({ inputData }) => {
        return { output: `Processed: ${inputData.input}` }
    },
})

const myWorkflow = createWorkflow({
    id: 'my-workflow',
    inputSchema: z.object({ text: z.string() }),
    outputSchema: z.object({ result: z.string() }),
})
    .then(myStep)
    .commit()

const result = await myWorkflow.execute({ text: 'Hello' })

Creating a Network

import { Agent } from '@mastra/core/agent'
import { googleAI3 } from '@mastra/config/google'
import { pgMemory } from '@mastra/config/pg-storage'

const myNetwork = new Agent({
    id: 'my-network',
    name: 'My Agent Network',
    description: 'Routes requests to appropriate agents',
    instructions: ({ agents }) => `
    You coordinate between multiple agents:
    ${Object.keys(agents).join(', ')}
    
    Route requests to the most appropriate agent.
  `,
    model: googleAI3,
    memory: pgMemory,
    agents: {
        // Add your agents here
    },
})

const result = await myNetwork.generate('Research this topic')

UI Components

AgentStack includes 49 UI components for building AI interfaces:

Chat Components: Real-time conversation interfaces
Reasoning Components: Thought process visualization
Canvas Components: Interactive agent workspaces
Data Visualization: Charts and analytics displays

Features Overview

Component	Count	Purpose

Next Steps

Explore Agents - Check /src/mastra/agents/ for available agents
Configure Tools - Set up your API keys in .env
Try Workflows - Start with /src/mastra/workflows/
Build Networks - Combine agents in /src/mastra/networks/
Create UI - Use components from /src/components/ai-elements/

Development

Run tests:

npm test

Build for production:

npm run build
npm run start

Resources

Documentation: Comprehensive guides in /docs/
API Reference: Tool and agent specifications
Examples: Working implementations in /examples/
Changelog: Latest updates and features

Framework Version: 1.0.0 | Last Updated: November 2025