Progress log #2

[TheHunterDog]

Last weekend, I attended the Y Combinator event in Paris. I learned some new things and met amazing, ambitious people. This really boosted my motivation and improved my ambitions for this project!

My ambitions with this application are similar to before, but the goal is to have access to all the AI’s in the application. So, an integration with Antropic, Gemini, ChatGPT, and Grok APIs would be amazing. However, the application won’t provide API keys; these must be provided.

This weekend, I made some progress, although I did cheat a bit. I’m using hardcoded values for numerous items, such as the base URL, bearer token, model, and chat ID.

I’ll change these items today and also do some code cleaning. If everything goes well, there will be a TestFlight next week!

I did face some challenges, which I partially resolved:

• The assistant chat messages don’t show up in the web interface (the message is present in the fetch of the chat history).
• The whole API of OpenWebUI works a bit inefficiently in my eyes, which will also affect the application’s performance. (On each chat message, the whole chat history is sent twice to the API.)
• The response from the assistant is capable of streaming, but this will be added in the future. I’m focusing on the basics first and adding features like streaming, rag, and image generation in the future.

What’s next?

• Login screen to provide the base URL and credentials.
• Show the chats or create a new chat.
• Code cleaning.
• Show models that could be used.

Important notice:

Currently, there is no Patreon to support this project.

If you have any features you’d like to see, please leave a comment or create an issue. This will help us understand what you want and prioritise our work. If no issues are made, I’ll focus on other projects that may not be as valuable to you.