- The Homebase AI
- Posts
- 1996 All Over Again? The Framework Shaping the AI Agent Era
1996 All Over Again? The Framework Shaping the AI Agent Era
Learn Mastra’s 5 core principles — from orchestration to continuous learning — and see how developers deploy agents into real workflows today.
David Stepania
August 28, 2025
🎤 Full Interview: Sam Bhagwat, Founder & CEO @ Mastra
"This is bigger than anything we’ve seen—this feels like 1996 all over again, and the winners will be the ones who ride the wave, not get washed away. "
Founder Story & Vision
Who they are & what they’re building
Sam Bhagwat is a veteran open-source JavaScript engineer—previously co-founder of Gatsby—and today he’s on a mission to simplify AI agent development with Mastra, a TypeScript framework that orchestrates data triggers, workflows, specialized agents, and deployments into real-world applications.
Why now & what’s the big bet
After his first startup, Gatsby, an open-source, React-based framework that is still used by over 200,000 websites, allowed developers to build 1,000x faster on Netlify, Sam saw a similar inflection point in AI agents: businesses are drowning in repetitive tasks, and developers need battle-tested orchestration tools to unlock productivity. Mastra’s bold bet is that JavaScript developers will dominate this era, but only if they adopt the right frameworks now—or risk being automated out.
Principles of Building Scalable AI Agents:
Mastra’s framework is built on five core principles: orchestration, developer-first design, production reliability, composable workflows, and continuous learning. Together, they ensure AI agents aren’t just prototypes—but production-ready systems that scale with real business needs.
🧩 Real-World Use Cases
How it works in the wild
PLAID Japan uses Mastra to automatically spin up financial data agents in Google Cloud, cutting integration time from weeks to days.
An aerospace supplier ingests 100-page spec PDFs to generate and validate circuit diagrams in under five minutes—tasks that once took engineers hours.
A veterinary clinic routes exam transcripts through agents to auto-create treatment plans and follow-up reminders, boosting staff throughput by 40%.
What you’ll learn:
How Sam navigated the shift from cloud-first to agent-first infrastructure
Mastra’s framework for error-tolerant task routing and context management
Tactics behind Mastra’s 2–3× dev productivity gains using code-review bots and multi-agent workflows
Lessons on hiring, community-driven growth, and commercial open-source go-to-market
How Mastra tackles the “paperwork-intensive” problem across industries
Some Takeaways:
Adopt agent orchestration early to avoid becoming a legacy developer
Use open-source frameworks as the front door to your cloud product
Measure success by productivity gains, not just POC completion
Leverage code-review bots and autonomous agents for continuous delivery
Start with industry-specific workflows (e.g., claims processing, circuit design) before generalizing
Listen on Spotify | Apple Podcasts
Watch on YouTube
In this episode, we cover:
00:00 – Introduction
02:15 – Lessons from Gatsby
07:40 – Spotting the AI Inflection Point
12:30 – Why Orchestration Matters
18:05 – Mastra’s Bold Bet on JavaScript Developers
24:50 – Building a B2D Company
31:15 – Founder Insights & Hard Choices
37:20 – From Framework to Adoption
44:10 – Looking Ahead
49:00 – Closing Reflections
For inquiries about sponsoring the podcast, email [email protected]
Referenced in the Episode:
Mastra Orchestration Framework
Gatsby (React)
PLAID Japan
Surge (data provider)
“Principles of Building AI Agents” (Mastra book)
Find Case Studies of all other episodes here.
Reply