SpendLens AI
SpendLens AI is a hosted AI cost optimization and observability product from Rangkash PTY LTD. It connects to OpenAI and Anthropic accounts to analyze AI spending and identify which workload, model, or API key is driving costs. The product positions itself as model right-sizing for AI, similar to EC2 rightsizing, and focuses on converting provider billing totals into actionable cost drivers. After identifying high-cost workloads, SpendLens AI can compare lower-cost model candidates and run quality validation before production changes. Candidate calls run in the customer's staging application with their own keys, while optional quality evaluation runs separately on SpendLens infrastructure using redacted templates and safe generated data. The application continues calling OpenAI and Anthropic directly; SpendLens AI is not a proxy, gateway, or model provider. It is aimed at engineering and product teams that need to control AI infrastructure expenses without rewriting applications or adding request latency. The homepage advertises no SDK requirement to start, no proxy, and no production change required. It provides sample reports and savings estimates, and supports connecting OpenAI or importing Anthropic data.
- Category
- Observability
- Pricing
- Free
Problem solved
Provider dashboards typically show total AI spend but not which project, workload, model, or API key caused it. SpendLens AI attributes OpenAI and Anthropic costs to specific workloads and identifies lower-cost model candidates, helping teams find and validate savings before changing production.
Use cases
- Attribute OpenAI and Anthropic spending to workloads, models, or API keys
- Identify the highest-cost AI workload in an application
- Compare a production model against lower-cost alternatives
- Run automated quality validation before switching models in production
- Estimate potential monthly savings from model right-sizing
- Monitor AI FinOps for engineering and product teams
Audience
- Developers
- ML Engineers
- Product Managers
- Enterprise
- Startups
Pros
- No SDK or proxy required to start
- Attributes costs to workloads, models, and API keys rather than just showing a total
- Validates cheaper model quality before production migration
- Does not become a gateway or add request latency
- Supports both OpenAI and Anthropic cost analysis
Cons
- Limited to OpenAI and Anthropic providers
- Requires connecting or importing provider account data
- Candidate model tests run in the customer's staging application, requiring setup
- Hosted quality evaluation uses redacted templates and generated data, not full production traffic
- No public GitHub repository or open-source code available
Official website: https://spendlensai.dev/
Canonical LambdaBase page: https://www.lambdabase.com/tools/spendlens-ai