This is a fantastic piece of software that allows you to use your own API keys (or optionally, use their cloud service). It is efficient, works well and syncs with the mobile app.The only drawback is that it doesn’t let you upload arbitrary files for context like the regular ChatGPT or Claude apps do. I will like the developer to add support for it.

As a third-party client that utilizes ChatGPT API, OpenCat is undoubtedly a practical and recommendable application. As a user who is interested in conversational bots, I found this application on the App Store and realized it is both user-friendly and highly customizable.OpenCat enables you to easily integrate with ChatGPT API, powered by natural language processing technology, and launch your own chatbot. You can also customize its domains, entities, intents and other components, making it more suitable for your needs. This is very useful because it allows better control over your bot's behavior and makes it more human-like.Moreover, I appreciate OpenCat's interface design. Its interface is straightforward, easy to use, and helps you create and manage multiple chatbots with ease. I can effortlessly add my own dialogs to my bot, modify them when needed, and make it more engaging to interact with my users.In conclusion, I really like OpenCat and I think it is an extremely practical, user-friendly, and highly customizable application. Whether you're interested in conversational bots or you want a customized chatbot to meet your specific needs, I highly recommend using OpenCat.

This application stands out as one of the most visually appealing AI chat platforms available, and its array of features, including API integrations with various AI service providers and multi-device synchronization, warrants a 4.5-star rating from me. However, there are some areas that need improvement:1. Folders: The lack of folders for organizing chat threads makes it challenging to manage even a moderate number of conversations, let alone hundreds.2. Model Accessibility: The "default model" selection doesn't reflect the models available through API integrations, which hinders users from utilizing services like Ollama or Groq within the app's extension or menu bar interface.3. Email Integration: There appears to be a significant issue with the email reading functionality; the AI response suggests manually inputting an email for a reply, which is problematic since I would expect the extension to pass the email content to the model. This was my primary reason for considering the purchase.4. Keyboard extension missing: The "Keyboard extension" option isn't available in the preferences menu as advertised on the website and listed among premium features, making it difficult to use within other applications.5. Communication: There's a notable absence of direct communication channels with the development team — no forum, email address, roadmap, or feedback form — which necessitates detailed reviews like this one. I hope they introduce more user-friendly ways for communication soon.6. Family Sharing: Enabling Family Sharing would be beneficial for many users, including myself, enhancing collaboration and accessibility within households.The app's overall design and functionality are commendable, but these points require attention to improve the user experience significantly.

I run my own AI models and I was looking for an app to use across my Apple devices that would give me a ChatGPT-like experience with my own AI. The interface for this app is *almost there*.But there are other things that look like maybe they were started awhile ago but never completed? Examples:* MCP Servers - stdio and sse are presented in the interface. But stdio isn’t supported on Apple devices (so why even present the option in the interface?) and sse is a deprecated protocol. Streaming MCP is the way to go now but it’s not supported in this app. So technically, yes, there’s MCP. But outside of a few edge cases you can’t really use it.* Search - There’s a search button in the chat interface but options are greyed out and written in a language I can’t read. There’s no way to configure this. So it’s another interface option that looks like maybe it was started but never finished.We really need some tools *like* this and I really encourage the developer to see this through. Polish it up. There’s potential for greatness here, but it’s still unrealized.