Show HN: Frugal Tokens – explore costs and usage across coding agents

Re:Show HN: Frugal Tokens – explore costs and usage across c... by joshstrange on Wednesday August 19, 2026 @01:56PM (#49364877)

I can highly recommend https://agentsview.io/ for a way to dig into your LLM usage. It reads your local conversation files and provides a ton of visualizations as well as the ability to dig into each conversation and see all the tool calls/etc. It's incredibly powerful.

[ Parent | on HN ]

Re:Show HN: Frugal Tokens – explore costs and usage across c... by dpc94 on Wednesday August 19, 2026 @02:07PM (#49365000)

Looks interesting. One of the angles that I approached while building this was focusing on cache misses. Not sure if this shows much info about this category. My tool is also local only + detecting changes and syncing into a sqlite db

[ Parent | on HN ]

Re:Show HN: Frugal Tokens – explore costs and usage across c... by joshstrange on Wednesday August 19, 2026 @02:59PM (#49365760)

Yes, it does show cached vs not:

- https://cs.joshstrange.com/jzwC4Z9y

- https://cs.joshstrange.com/nynGc5V9

[ Parent | on HN ]

Re:Show HN: Frugal Tokens – explore costs and usage across c... by dpc94 on Thursday August 20, 2026 @09:05AM (#49374105)

yeah i tried to do even deeper and classify the types of misses and their monetary cost. for example, I was using Fable at work, and a single TTL miss (responding an hour and half later), cost $6 for one message

[ Parent | on HN ]