Narev provides FinOps infrastructure for Large Language Model usage, offering real-time token pricing data and billing components that help engineering teams track, control, and optimize AI spend across multiple providers. The platform supports REST, MCP, and Skills-based access to pricing for models from 20+ providers, including Anthropic, OpenAI, and Google. Narev's tooling also includes credit wallets, a policy engine, and auditing ledgers for organizations scaling AI costs into product economics.
Funding
Funding not disclosed
Founders
Product
Problem
Engineering teams and organizations face significant difficulty managing the rapidly evolving and highly variable costs of Large Language Model (LLM) usage. With pricing changing frequently across a fragmented landscape of over 20 providers and thousands of models, teams lack real-time visibility and tooling to track, govern, and optimize AI spending effectively.
Solution
Narev provides an infrastructure platform for LLM FinOps, offering real-time token pricing data and comprehensive billing components to help teams control AI costs. The service exposes a REST API, MCP server, and Skills interfaces to deliver current pricing for models across major providers, including Amazon Bedrock, Anthropic, OpenAI, and Google. Beyond pricing lookups, Narev offers drop-in billing components like top-up mechanisms, usage tracking, and model selection widgets. For teams scaling AI into product economics, the Enterprise tier adds credit wallets, a governance policy engine, an auditable usage ledger, and itemized billing to serve as a complete financial operating layer for LLM infrastructure.
Target Audience
Narev is built for engineering teams at companies of all sizes, from individual developers to large enterprises, who need to embed LLM usage into their product economics and require granular tracking, budgeting, and cost governance across multiple AI providers.
Features
- Real-time pricing API covering over 20 LLM providers, including Anthropic, OpenAI, Amazon Bedrock, DeepSeek, and NVIDIA, with per-model input/output token costs
- Drop-in billing components for user top-ups, usage tracking, model selection, and pricing display
- Credit wallet system with grants, expiry rules, priority consumption, and debt recovery mechanisms
- Policy engine allowing configurable margins, rounding rules, BYOK policies, and free-tier behavior
- Auditable usage ledger with reconciliation, retry, and replay support for accurate financial tracking
- Access via multiple interfaces: REST API, MCP (Model Context Protocol), and Skills for flexible integration