OpenAI spend shows up as a single line item. Attribute decomposes it. The eBPF sensor reads API calls at runtime, captures model identifiers and request metadata, and attributes cost to the exact microservice, customer, or feature generating the requests. The first AI cost visibility capability Attribute shipped.
WHAT YOU GET
→ Model-level breakdown: GPT-4o, GPT-4, GPT-3.5, o1
→ Prompt vs completion token cost split
→ Cost attribution to calling service and downstream customer
→ LLM gateway decomposition support (OpenAI-native)
→ Spend trend and anomaly detection
BEST FOR
Any SaaS company with OpenAI in production. Particularly valuable for teams seeing unexplained AI cost growth.