Backstory: after spending the better part of a decade building Gatsby, we joined Netlify. And earlier this year, we began building an AI-powered CRM called Kepler.
Building Kepler, we tried a few different approaches for the AI bits. But nothing felt....right.
*Domain-complete.* We generally settle on a framework that has opinions about all the key jobs-to-be-done in a domain. Next.js, Kubernetes, Terraform, etc.
Mastra has top-level nouns for all the key concerns in AI engineering: LLMs, agents, workflows, RAG, observability.
Comments
Loved the langgraph concept but can’t stand writing another line of python.
Building Kepler, we tried a few different approaches for the AI bits. But nothing felt....right.
It's built for product developers: frontend, fullstack, backend, the kinds of folks that had fun toying on their
Gatsby blogs years and years ago.
*Typescript-first, Typescript-only.* Typescript is the language of product development.
Pythonic AI frameworks have JS/TS ports, but they're an afterthought, with a fraction of the docs/examples.
Mastra includes a workflow graph with branching, waiting for human input, embedding other workflows, error handling, etc.
Mastra has top-level nouns for all the key concerns in AI engineering: LLMs, agents, workflows, RAG, observability.
AI frameworks have too many, poorly organized APIs.