Getting Started with Mastra (TypeScript Agents)

Mastra is a TypeScript agent framework: workflows, tools, memory, and a local Studio. Home: mastra.ai. Docs: mastra.ai/docs. Scaffold: npm create mastra@latest.

This is the TS contrast to Python stacks (smolagents, PydanticAI). After work, use it if your day job is Node/React—not as a replacement for learning Python.

Who this is for

Frontend or Node engineers who want an agent in TypeScript they can git. If you only have Python tonight, skip this and stay on a Python framework.

Overview: Top 10 AI agent tools for after-work Python.

Steps to start tonight

1. Create a tiny project

Need Node.js 20+. In an empty folder: npm create mastra@latest. Accept defaults. Open the generated src/ agents file.

2. Put the API key in .env

Copy .env.example. One provider key. Never commit .env.

3. Keep one agent, one tool

Start with a weather or docs-lookup tool that hits a public JSON URL you control. Delete extra sample agents if they confuse you.

4. Run Studio locally

Follow the printed command (often npm run dev). Studio is commonly http://localhost:4111. Chat once. Confirm the tool ran.

5. Stop before deploy

Do not push to a public cloud or add Slack write tools tonight.

Tiny exercise (20–30 minutes)

Write a tool that returns { name, officialDocs } for a library name. Ask the agent for Mastra vs one Python agent library. Require URLs. Compare the TypeScript types to a Python dict.

Pitfalls

• Treating Mastra as the default after-work Python path.
• Shipping with sample keys still in git.
• Five workflows and a vector store on night one.
• Mixing this with coding agents (Amp, Cursor)—Mastra is a library, not an IDE.

Why this fits after work

You feel typed agents in one evening and know when TS is worth it versus Python.

Next step

Keep Python as the after-work default with Intro to Python Programming.