Fake limits that still did not fix Kimi
**Kimi / Moonshot has serious UX and transparency problems — here's a breakdown for anyone considering paying**
I've been using Moonshot's Kimi API and CLI tools for a while now, and I want to share some genuine issues I ran into — not to bash the product, but because I think it has real potential and these problems are holding it back.
---
**🔢 Usage limits are scattered across multiple places**
There is no single, clear dashboard where you can see your current usage and limits at a glance. Instead, you're hunting across different pages and panels to piece together where you stand. This is a fundamental UX failure. One place. That's all it takes.
---
**💸 Pricing transparency and refund policy**
I paid ~$100 expecting a comparable experience to tools like Claude Code — same category, similar promise. The output quality and reliability didn't match that price point for my use case. That alone I could accept — but here's the real issue: the 14-day no-refund policy (which applies under EU consumer standards) is not prominently communicated upfront. You only notice it toward the end of the process, after you've already committed. That's not okay.
---
**🔐 Basic account security features are missing**
The web app has no password reset button and no "log out all devices" option. These are not advanced features — they're table stakes for any SaaS product in 2025. This is a real concern from a security standpoint.
---
**🍝 Overall UI/UX feels unfinished**
The interface feels like it was built in silos — there's no coherent flow between sections. Navigating between billing, limits, settings, and the actual product feels disjointed. For a product positioning itself as a premium tier among affordable AI tools, this matters.
---
**What I'd love to see:**
- A single unified usage/limits dashboard
- Clear refund policy disclosed at checkout
- Password reset + session management on the web
- A UX pass that connects the dots between key sections
I actually built a full CLI suite around this product because the core tech is promising. That's what makes the rough edges more frustrating — the foundation is there. Hopefully this gets to someone on the team or helps others set expectations before committing.
Anyone else running into these issues?
