The vision and direction of Nily AI are quite impressive. With its aspirations to unify various AI models under one roof, it holds the potential to be an exceptional tool—it could easily reach 5 stars if everything worked as intended.

The Challenges:
Inconsistent Availability of Models:

One of the most significant issues is the inconsistent availability of large language models (LLMs) on the platform. Often, models like Grok 4 are either partially functional or entirely unavailable, whereas ChatGPT stands out as consistently reliable.

I tested grok 4 during the making of this review to prove my point & it wasn't available. I asked "What time is it right now?" and the LLM responded with "｛"type"："answer"l"answer"："抱歉，作为AI助手，我没有实时访问当前时间的权限。时间因您的位置和设备而异，请查看您的设备时钟或使用本地时间服务来获取准确的当前时间。如果您有特定时区的要求，我可以帮助计算相对时间！"｝"

Not sure what that means but I'm just happy it responded at all as sometimes the LLM (not exclusive just GROK) won't respond at all. Other times it responds with random answers having nothing to do with the conversation.

Unreliable Image Generation Models:

The image generation tools, such as Flux and ReCraft, work sporadically. It seems that server capacity might be an issue, leading to frequent downtimes or access problems during peak times.
Feature Limitations:

Nily AI lacks certain critical features available on other platforms. For instance, it doesn’t support uploading reference images or allowing prompts longer than 200 characters, which limits its creative scope. Some of the image gen tools with Nily do allow references to be uploaded, but not all of them and not on every page. Image Gen tab has more features available than Image editor, even when using the same LLM.
Conclusion:
While Nily AI promises a comprehensive solution by bringing multiple tools together, these reliability and feature issues make it a work-in-progress rather than the finished article it aims to be. Users may find it challenging to fully rely on it for their daily AI needs, given its current limitations.

In summary, Nily AI showcases potential but requires improvements in model stability and feature breadth to expand its utility and effectiveness. I hope the team can focus on enhancing these areas to unlock the platform's full potential.

I want to end with making sure I state that when/if this product works to it's fullest compacity, you will regret not buying the LTD. It is really a steal even without everything working. When everything works, it won't make sense to be this inexpensive. If they add agentic agents or an agent to the mix where you can master prompt it, this platform will be the best I've ever seen within the AI space.