No need to emphasize "by ChatGPT" now

Dark mode on opening nav items

Add dark mode toggle (#46)

Lesser update frequency to use less DB compute hours

Fix unescaped html tags from llama model

Handle groq TPM limit

Try qwen model

Support reading/writing the URL hash string

Avoid checking whole list everytime to reduce transfer cost to DB

A more prominent comments icon

Halve update frequency to reduce traffic to the database.

fa-heart -> fa-chevron-up

Print content log earlier in case api call throws

Add stepfun model

Upgrade pdfminer to avoid hanging when parsing large PDFs