Reviewers see MGX as a promising AI dev workflow that helps people turn ideas into webpages, code, UI structure, and early project plans quickly, with a clean interface and especially strong appeal for solo builders, prototyping, and beginners. But praise is tempered by repeated concerns: suggestions can feel generic, editing UI is awkward, and context handling, reliability, and integrations still need work for serious real-world use. A few users report much harsher experiences, citing defects, broken links, wasted credits, and poor support.
Atoms
Thank you for sharing this detailed feedback. We're very sorry for the frustrating experience you had, especially after investing so much time and credits. The "fixing spiral" you described is exactly the kind of problem we want to eliminate.
Your feedback on the high credit cost during iterative fixes is incredibly valuable, as it's an area we've focused heavily on. In a recent update (V1.1.3), we specifically optimized our token usage for multi-round incremental development, significantly reducing costs for scenarios just like yours.
Additionally, we introduced an "Engineer Mode" (V1.1.9) which is designed for more focused tasks like debugging, making the process both faster and more credit-efficient.
We know this doesn't change your past experience, but we are committed to making MGX more reliable and cost-effective, and we appreciate you taking the time to highlight these critical issues.