Can you list all unique features this project provides which make this better than openwebui ?

[Greatz08]

I have seen numerous similar projects on github like this one, and i ask the same question from those projects main devs. I want to understand what all unique features this project provides when compared against extremely powerful and capable project - openwebui.

[monayman]

It is garbage every task I have it do , each individual chat corrupts itself with errors that aren't fixable without starting a new chat and if you have that new chat investigate it corrupts that chat as well 🤣. Then you have the ongoing issue of censorship and restrictions trying to force prompts to bypass even if what your doing isn't bad at all. It's a headache all around. It's way easier to take the offline model remove all restrictions (that takes forever) and run it on your own infrastructure to be unrestricted but still give it access to search and rootFS. Sure deepchat can be helpful if you only use it for the basics and I mean the basics but for someone who is experienced , into cyber security or any type of higher level of knowledge this is pointless. Hope they fix these ridiculous issues but I doubt it'll happen when it's intentional.

[zerob13]

It is garbage every task I have it do , each individual chat corrupts itself with errors that aren't fixable without starting a new chat and if you have that new chat investigate it corrupts that chat as well 🤣. Then you have the ongoing issue of censorship and restrictions trying to force prompts to bypass even if what your doing isn't bad at all. It's a headache all around. It's way easier to take the offline model remove all restrictions (that takes forever) and run it on your own infrastructure to be unrestricted but still give it access to search and rootFS. Sure deepchat can be helpful if you only use it for the basics and I mean the basics but for someone who is experienced , into cyber security or any type of higher level of knowledge this is pointless. Hope they fix these ridiculous issues but I doubt it'll happen when it's intentional.

Could you provide some screenshots or logs regarding the conversation corruption issue? That would help me better understand and diagnose the problem.

As for new conversations, it’s true that they currently do not support using past conversations as context. However, they should not be getting corrupted, so that part is quite strange.

Regarding censorship or restrictions, DeepChat is simply a chatbot client and does not include any such mechanisms. The mention of offline models is also confusing, as DeepChat does not provide any models itself.

Let me emphasize once more: DeepChat is only a desktop client for accessing the models you configure. The behavior of the model itself is not implemented by this project, so please avoid mixing them together.

[zerob13]

I have seen numerous similar projects on github like this one, and i ask the same question from those projects main devs. I want to understand what all unique features this project provides when compared against extremely powerful and capable project - openwebui.

---

Although many of these projects share the same core concept—calling an LLM API for chat—there are several reasons why this project attracted me and why I chose to contribute:

1. A full-featured, commercially friendly ChatBot
   I want to build a feature-complete chatbot with a business-friendly license. I’m a big fan of the Apache 2.0 license because it’s permissive for commercial use while still offering some protection for open-source authors. I believe a ChatBot should be a fundamental tool/infrastructure, and therefore it should have a relatively free and open version. Thanks to this licensing choice, our project has already attracted companies willing to pay for customization. It’s not a huge amount yet, but it’s a very healthy start for the long-term growth of the project.

2. A preference for native-style desktop apps
   I personally prefer desktop applications. I don’t like putting everything into the browser. (Even though DeepChat is also based on Electron—that’s simply due to my own technical limitations. If I had the capability, I would probably build it using native technologies like C# or Swift.) A desktop app requires no server deployment—just download, install, configure, and you’re ready to use. For people like me, that experience is ideal.

3. More aggressive exploration of new LLM-related technologies
   We tend to adopt and experiment with new technologies earlier. For example, we added MCP support back in March of this year (add mcp support for deepchat #129), and recently we added ACP support as well ([Feature] Add ACP Support #1100). These features are extremely helpful for many agent/LLM developers as debugging and experimentation tools. I’ve received feedback from developers of other projects saying that these early explorations were valuable references—and that means a lot to me.

4. A love for the Vue ecosystem
   Most similar projects are built with React, and OpenWebUI uses Svelte. These are all great ecosystems, but personally I feel that Vue is better suited for this type of application—so I wanted to build something with it and see how far it can go. During development, we also open-sourced our streaming Markdown rendering component https://github.com/Simon-He95/vue-markdown-renderer . Thanks to Vue’s design, its streaming Markdown performance is noticeably better than the commonly used Streamdown component in React.

5. For joy—writing software is fun :)
   At the end of the day, programming itself is enjoyable. That’s reason enough, isn’t it? :P