cloudshrink.com

What It Is

Cloudshrink is not a traditional marketing or SEO tool, but a cost intelligence and unit economics platform for AI products. It describes itself as a “managed FinOps team — not a dashboard.” Its core goal is to attribute every LLM API call to customers, features, teams, projects, and models, helping teams answer management-level questions such as “What is the AI cost per customer?”, “Which AI features have positive gross margins?”, and “Why did AI spend suddenly rise last month?”

Core Features and Data Sources

Based on the available text, Cloudshrink supports connections to providers including OpenAI, Anthropic, Google Gemini, AWS Bedrock, Azure OpenAI, Mistral, DeepSeek, and Groq, importing cost and usage data in read-only mode. Its documentation also shows SDK-based instrumentation, with support for Python and npm. It can collect fields such as provider, model, input/output tokens, cached tokens, project, actor, env, and cost_usd. Reported metrics include cost per customer, cost per feature, cost per AI interaction, product-level margins, unattributed spend, project/actor/model analysis, and cost change detection.

Pricing and Service

The Platform plan starts at $500/month and includes cost attribution, unit economics dashboards, support for all providers, unlimited projects/teams/features, and spike detection alerts. Platform + Advisory is custom-priced and adds hands-on audit, monthly unit economics reviews, board-ready reporting, and a dedicated customer success manager. The page states that pricing varies with the scale of managed AI spend, and that most customers pay around 2% of tracked LLM spend. In addition, qualified companies with over $10K/month in LLM API spend can apply for a free audit.

Pros and Cons

Its strengths are a clear positioning around the problems of runaway AI costs and opaque unit economics, along with support for multiple LLM providers. Compared with a pure observability dashboard, it also emphasizes monthly reviews, proactive recommendations, and board-readable reports. The downsides are its relatively high entry threshold: the $500/month starting price and the $10K/month requirement for a free audit are not ideal for small experimental projects. The crawled content also does not specify key procurement details such as security/compliance certifications, SLA, or payment methods.

Who It’s For and Access from China

Cloudshrink is best suited to AI SaaS companies, AI-feature product teams, CTOs, engineering leaders, and finance teams that already have relatively high LLM bills, especially for pricing, gross margin analysis, model switching, and anomaly investigation. The crawled text does not provide information about access from China, so its status is unknown; payment methods are also not disclosed. Possible alternatives include Langfuse, Helicone, OpenMeter, Vantage, and CloudZero, though their capabilities should be compared specifically by the depth of AI cost attribution they offer.