Keep track of token usage #pliik

find out how to track used amount of tokens #linqmeup

increment messages sent in db to track usage #therapistai

register conversions to see usage per user, tokens used and cost #linqmeup

Dynamically choose how many recent messages to include for context, based on total token lengths. Instead of just taking the recent 10 messages like I did before, which meant sometimes I wasn't using as much context as I had available, and other times I was exceeding the limit of tokens which broke ChatGPT #marcbot

log token usage per request #linqmeup

Add token stats, cost price to admin accounts #veda

notification token optimization on gocapture #meratek

#pn Count the number of messages the bot receives

Be able to calculate messages number based on total characters given #stenograph

When a message doesn't get process properly due using too many tokens, keep retrying it, each time with one fewer message #marcbot

add #hoodmaps usage stats to daily Telegram bot

#pn implement basic metrics to count api requests, messages scheduled and so on

Create new Twitter optimization GPT

Prompt ChatGPT with custom embeddings #life

measure click event of token generation page #penelopeai

write code with GPT4 to fix prompt weighting from ((model)) to (model)++ #photoai

add rough API cost tracking in db with price per model training, price per outpaint job, price per upscale saved and incremented in db #avatarai

GPT4 for custom texts for reports #indiehacking

play with system prompts and chatGPT API to see if we can replace fine-tuned model #reva