Cloudshrink attributes every LLM dollar to a project, a person, and a model. Go from spend spike to root cause in under 60 seconds—across every provider you use.
Attribute costs across every provider. One instrumentation, one dashboard.
Every dollar traced to a project
Tag each LLM call with a project name. See exactly which projects drive your spend, and drill down to the actors and models behind each one.
Know who's spending what
Whether it's a developer, a service account, or a CI pipeline — every token has an owner. Spot the top spenders and understand their usage patterns.
From spike to culprit in minutes
Automatic period-over-period comparisons flag any project, actor, or model with a >20% cost increase. See what changed, who changed it, and when.
No more mystery line items
Cloudshrink tracks events missing project or actor metadata and shows you exactly how much spend can't be attributed. Fix your instrumentation, close the gaps.
Pick a project from our demo data to see how attribution breaks down spend by actor and model.
Swap one import. Our SDK wraps OpenAI, Anthropic, and Google clients. Tokens, costs, and latency are captured automatically.
Add project and actor to each call as metadata. Two fields, one line. Everything else is inferred.
Open the dashboard. Every dollar is traced to a project, a person, and a model. Spikes are flagged automatically.
Have the answer before the question. Project, person, model, timestamp—all in one click.
See the Live Demo