I've updated the Function to display Token Usage & Costs
well.. About 5 months ago, I had some free time and built an OWUI extension to keep track of my session size and avoid accidentally exceeding the context window limit.
I always also had a concern that session performance might degrade once it exceeds 30-50% capacity (this is something I've noticed during my own testing with different models).
So, I put together a simple extension, published it, and completely forgot about it.
Recently, I discovered that the extension has been downloaded over 3,700 times. Whaaat? I was quite surprised, as I assumed such a niche tool wouldn't have much appeal.
So, I decided to update it.
Here is a brief overview of the plugin: Shows token counts, reasoning/cached breakdowns, context-window utilization, generation time, tokens/sec and message/chat cost below each AI response. Reads provider-reported usage across OpenAI (Chat & Responses API), Anthropic, Gemini, Ollama and llama.cpp, with an optional tiktoken fallback. Cost is shown when your provider/proxy reports it (OpenRouter, LiteLLM) and can optionally be estimated from models.dev prices.
https://openwebui.com/posts/token_usage_display_a94ea72f
Just wanted to share this in case anyone finds it useful. No rocket science or smart AI things, just a simple QoL improvement =)