DEMO — mock data, not real spend.  back to nomiraai.com
← all projects

NOMIRA · project: data-pipeline

C--Users-team-Documents-work-data-pipeline  ·  70 sessions · 42,430 calls
Today7d30dAll
API-equivalent value
$12.22
Share of your AI value
18%of $66.61 total
Sub-agents (Task)
17~$28.70
Bash calls
3199~$934.04

Recommendations (rough savings — attribution is heuristic)

File read 75× — pin / summarize / .claudeignore
C:/Users/team/Documents/work/data-pipeline/src/data-pipeline/store.ts
$167.89est. savings
File read 102× — pin / summarize / .claudeignore
C:/Users/team/Documents/work/data-pipeline/src/data-pipeline/router.ts
$140.29est. savings
File read 161× — pin / summarize / .claudeignore
C:/Users/team/Documents/work/data-pipeline/src/data-pipeline/handler.ts
$137.77est. savings
Bash dominates (3199 calls)
Pipe heavy outputs through `head`/`grep`/`wc` to keep context lean.
$186.81est. savings
17 sub-agent (Task) runs
Batch related tasks, or route sub-agents to a cheaper model when adequate.
$8.61est. savings

Top tools by attributed cost

Bash
$934.043199
Edit
$650.181752
Read
$470.80938
Grep
$242.48314
Write
$93.69122
Task
$28.7017

Top sessions (by API-equiv value)

SessionCallsCost
sess00aa…606$60.11
sess01aa…606$66.13
sess02aa…606$72.14
sess03aa…606$78.15
sess04aa…606$84.16
sess05aa…606$90.17

Most-repeated file reads (re-feeding the same file inflates cache reads)

FileReadsEst. costSave if 1×
C:/Users/team/Documents/work/data-pipeline/src/data-pipeline/store.ts75$170.16$167.89
C:/Users/team/Documents/work/data-pipeline/src/data-pipeline/router.ts102$141.68$140.29
C:/Users/team/Documents/work/data-pipeline/src/data-pipeline/handler.ts161$138.63$137.77
C:/Users/team/Documents/work/data-pipeline/src/lib/auth.ts160$107.34$106.67
C:/Users/team/Documents/work/data-pipeline/docs/data-pipeline/README.md114$51.57$51.12
Cache is 80% of this project's cost. Dollar figures are API-equivalent value (not a bill for subscription usage). Tool/file attribution is heuristic — cache cost is cumulative context, so per-item dollars are directional, not exact. File paths and tool inputs are read from local transcripts only; nothing is written to the team store.