CLI Reference

classify

Analyze the complexity of a prompt without running it.

$ infershrink classify "What is 2+2?"
Complexity: SIMPLE (score: 0.12)
Reasoning: Short length, no code blocks, basic arithmetic pattern.

$ infershrink classify "Implement a red-black tree with delete rebalancing"
Complexity: COMPLEX (score: 0.89)
Reasoning: Algorithmic request, requires structural code generation.

route

Simulate how a prompt would be routed given a specific model.

$ infershrink route "Hello world" --model gpt-4o
Original: gpt-4o
Routed:   gpt-4o-mini
Action:   Downgraded (saved ~95%)

$ infershrink route "Design a distributed consensus algorithm" --model gpt-4o
Original: gpt-4o
Routed:   gpt-4o
Action:   No change (task too complex)

stats

View global usage statistics and accumulated savings.

$ infershrink stats
Total Requests: 1,402
Routed Down:    982 (70.0%)
Original Cost:  $45.20
Actual Cost:    $12.40
Total Saved:    $32.80 (72.5%)

index

Create a semantic index from a directory of documents.

$ infershrink index ./docs
Indexing 12 files...
Created index at .infershrink/index.faiss (Size: 4.2MB)

clear

Reset local stats and cache.

$ infershrink clear
Stats cleared.