Tokenmeter

live
Export CSV
No transcripts found yet. Tokenmeter reads , and . Send one prompt in Claude Code, Codex or Copilot CLI and it will appear here within a few seconds. If your logs live elsewhere, set CLAUDE_CONFIG_DIR, CODEX_HOME or COPILOT_DB before starting the server.
How cost is calculated

Every model reply reports how many tokens of each type it used. Tokenmeter multiplies each type by that model's list price and adds them up:

cost = cached reads × read price + uncached input × input price + cache writes × write price + output × output price, all per million tokens

The types are priced very differently, so the token total alone does not predict cost. On Claude a cached read is about 2% of a fresh token, and output is 2 to 5 times a fresh token. Claude Code writes its cache with the 1-hour rate. OpenAI models do not charge for cache writes. The rates in use right now:

These are API list prices from pricing.json. If you are on a subscription you pay the plan price instead; the Spend tile shows that first and this API-equivalent figure below it.

Spend
–
Total tokens
–
Prompts
–
Cache hit rate
–
Output tokens
–
Model turns
–

Prompts Click a row for the full text, the cost math and a resume command · click Cost to see the most expensive first

Usage limits rolling windows, independent of the range filter

Active sessions context fill, last 24 hours

Usage over time

Token mix

Cost mix same tokens, weighted by price

Cache reads are cheap, cache writes and output are not. Thinking tokens are included in output.

By model

Cache misses

Cost per commit turns between one commit and the next, per repo

By project

Top sessions

Live feed