Feature request: Usage transparency — per-session/source token breakdown for paid accounts

Resolved 💬 2 comments Opened May 27, 2026 by emmahodgson77 Closed May 30, 2026

Problem

Claude Pro usage quota is consumed across multiple surfaces (claude.ai web, Claude Code CLI, remote agents/routines) but there is currently no way for a user to see what consumed their quota, when, or from which surface. When a user hits their limit unexpectedly, there is no log, no breakdown, and no way to investigate the cause.

This is a significant gap for paid subscribers. Being denied service with no record of how the quota was used is not acceptable for a metered product people pay for.

What I'm asking for

A usage log (in claude.ai settings or via API) that shows, at minimum:

  • Timestamp of each session/request
  • Surface (web conversation, Claude Code CLI, remote agent/routine, API)
  • Tokens consumed (input + output)
  • Session or conversation identifier

Why this matters

Every major metered cloud service (AWS, GCP, Azure) provides this as standard. Users should be able to audit their own usage, identify unexpected consumption, and make informed decisions about how they use the product. Without it, debugging unexpected quota exhaustion is impossible — including cases where usage originates from background agents or integrations the user may not immediately recognise.

View original on GitHub ↗

This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗