Mastra

building @mastra. author principles of building ai agents, the "most popular book in SF right now". prev cofounder @gatsbyjs

Mastra is an open-source TypeScript framework for building AI agents and workflows, aimed at product developers who would rather ship in their existing JavaScript stack than move to Python notebooks. The launch coincides with a $22M Series A led by Spark Capital, bringing total funding to $35M, and the debut of the Mastra platform. That platform includes Mastra Studio for evals, logs, traces, and datasets, Mastra Server for deploying agents and workflows, and Memory Gateway, an AI gateway with agent memory that works whether or not you use the rest of the Mastra stack. The company was founded in October 2024 by Sam Bhagwat (CEO), Abhi Aiyer (CTO), and Shane Thomas (CPO), the same team that previously built Gatsby.js, the React framework acquired by Netlify. Mastra reports customers including Brex, Indeed, and Marsh McLennan , with public examples ranging from a nationally advertised career counselor agent at Indeed and enterprise search used daily by more than 100,000 people at Marsh McLennan, to platform teams at MongoDB, Workday, and Salesforce automating DevOps and SRE work. The thesis behind the round is that as models have gotten better over the last 18 months, the harnesses around them have grown rather than shrunk, which is why Mastra has been shipping primitives like parallel tool calls, subagents, sandboxes, filesystems, and skills. For founders and engineering leaders evaluating where to build their next agent, this launch matters because it stakes out a credible TypeScript-native alternative to the Python-dominant agent tooling stack, backed by a team with a track record of turning developer frameworks into widely adopted defaults. Bhagwat is also the author of "Principles of Building AI Agents," which doubles as the conceptual playbook behind the product.

Tags

Comments (14)