Dashboard

All providers · last 30 days · sample workspace

$1,606

Spend, last 30 days

$1,659

Projected month run-rate · 66% of $2,500 budget

$54

Avg daily spend

$846/mo

Identified savings across 3 findings

gpt-4o$610 · 38%

claude-sonnet-5$482 · 30%

gpt-4o-mini$225 · 14%

claude-haiku-4-5$193 · 12%

o3-mini$96 · 6%

$412/mo

Right-size gpt-4o calls in Doc Summarizer

31% of gpt-4o requests are under 900 input tokens with simple summarization prompts. Routing them to gpt-4o-mini would cut cost ~94% per call.

$296/mo

Duplicate classifier calls in Support Bot

78% of intent-classification calls this week repeated identical inputs (612 of 785). Adding a response cache would eliminate nearly all of them.

$138/mo

Retry storm in Search / RAG

22% of calls are retries triggered by a malformed JSON output instruction. A stricter response schema would remove most retries.

This could be your spend — minus the waste.

Connect a read-only billing key. No SDK, no proxy, no card.