Mastra Roadmap: What's Next
This is a recording of a webinar from April 28, 2026. Choosing a framework for production agents is a bet on where that framework is headed. In this session, Alex and Shane (Mastra CPO and Co-founder) walk through Mastra's roadmap for the next three months - what's shipping, what's changing, and the technical bets driving those decisions. This isn't a tutorial. It's a direct look at where Mastra is going, designed for engineering leaders and technical decision-makers evaluating or already building on the framework. In this workshop, we're pulling back the curtain on our roadmap for the next three months, with a deep dive into coding and observability features that will change how you build and monitor agents. We'll start by exploring the evolution of the Agent harness. You've seen Mastra's harness capabilities; now our agents are becoming Claws. We'll talk through using Claw primitives—heartbeats, browser use, channels, and integrations, directly in Mastra. We'll talk about features that enable Long-Running Agents (LRAs): durable agents that can pause and resume work, background agents that run asynchronously, and dynamically created subagents that handle tasks as they arise. Then, we'll tackle the reality of scaling agents and observability features coming to Studio. We'll talk about features like auto-tagging, alerting, annotation, and review, building on top of the recent datasets and experiment features. 00:00 Intro 00:47 Why this session 04:10 Concierge and Mastra Start 05:12 Everything Mastra shipped this year 13:01 The Mastra platform: Studio, Server, Memory Gateway 14:08 Q: Why metrics needs an OLAP database 16:57 What's coming next on the framework 17:22 Long-running agents 20:54 Q: Dynamic sub-agents vs the supervisor pattern 22:23 Agents Love Mastra: A2A and a CLI for coding agents 25:18 Bringing OpenClaw primitives into Mastra 28:06 Q: Modes and the Mastra harness class 31:37 Observational memory retrieval 34:13 Mastra Code: now with a browser 34:56 Observability roadmap: annotations, review, alerting 36:46 Git workflows for the platform 38:09 Q: Mastra vs Langfuse, Braintrust, and Arize 39:54 Agent Builder: agents for the rest of your team 46:15 Q: The UI layer, AG-UI, and CopilotKit 48:50 Q: Agent Builder access and integrations 51:19 What Shane is most excited about 52:20 Follow us!

TypeScript BEATS Python when building AI Agents (Mastra's YC Journey)

Is Fine-Tuning Still Needed? LLMs, RAG, & LoRA

Monitor, Debug, and Evaluate Agents with Mastra Observability and Studio

The #1 GTM Engineer In The World | Jordan Crawford, Founder @ Blueprint GTM

How to Build AI Workflows with Mastra and AI SDK (Full Tutorial)

How I Code With AI Agents (Spec-Driven Development)

Andrej Karpathy: From Vibe Coding to Agentic Engineering w/ Stephanie Zhan

Why Every AI Agent Harness Turns Into a Claw - Mastra CEO Sam Bhagwat at Beyond Demos

Anthropic, OpenAI Should Not Be Allowed to IPO, Says Ed Zitron

How AI agents & Claude skills work (Clearly Explained)

Hermes Agent Fundamentals In 29 Minutes

Generative UI: Specs, Patterns, and the Protocols Behind Them (MCP Apps, A2UI, AG-UI)

Trump Slaps Fiery Tariff on Canada & SAVE Act Takes Voting Cues from Olive Garden | The Daily Show

Helios Is AMD’s First AI System To Rival Nvidia Vera Rubin — We Got An Exclusive, First Look

You Can Learn AI Agent Harness & Loop Engineering In 19 Min | LLM Ops, Eval, Tracing, RAG

Building AI Agents that actually work (Full Course)

Why Netflix is betting on systems thinkers—not specialists—in the AI era | Elizabeth Stone (CPTO)

Keynote: After the AI Hype – What’s Real, and What’s Next - Richard Campbell - 2026

Why Your AI Offer Isn't Selling, and How to Fix That
![Power Automate Beginner to Pro Tutorial [Full Course]](https://i.ytimg.com/vi/1p5kI7SYz4Q/hqdefault.jpg?sqp=-oaymwEjCNACELwBSFryq4qpAxUIARUAAAAAGAElAADIQj0AgKJDeAE=&rs=AOn4CLDIQUeJjCKSUU_QtkVwDZktEykVCg)
