Thanks for the feedback! Our mission with Mastra is to make sure users can get further faster with the standard primitives of AI engineering! We learned over the last year that while we want to be opinionated, users need to be able to eject and adapt Mastra however they want! We added server adapters so users didn’t have to deal with our opinions on esbuild, we made storage more configurable and extendable, and invested more into our workflows engine which will be the foundation for distributed execution in the future! We are always willing to hear feedback and welcome changes! Would love to know what you learn from the langchain comparison and happy to improve Mastra from it! Thanks!