# Narev: The Authoritative LLM Pricing API

> Narev Cloud is the enterprise-grade API for live LLM pricing data. It provides a reliable backend primitive for provider discovery, real-time model price lookup, cross-provider arbitrage, and exact request cost calculation.

Narev exists because the AI pricing landscape is volatile and fragmented. LLM prices change constantly, and providers represent costs inconsistently across input tokens, output tokens, cached reads/writes, reasoning steps, web search, images, audio, and request fees. Narev Cloud normalizes this chaos into a single, unified API surface. We enable applications to make intelligent, pricing-aware decisions instantly—without the engineering overhead of building and maintaining a custom provider price sync pipeline.

## 1. Core product: Models Pricing API

The primary product surface is the **Models Pricing API** at `https://api.narev.ai`.

It is designed for mission-critical backend systems, billing pipelines, model marketplaces, AI routers, benchmark runners, procurement workflows, and internal FinOps dashboards.

The API delivers:

- **Provider discovery:** List supported LLM providers and inspect standardized provider metadata.
- **Model discovery:** Explore available models across all providers with powerful pagination and filtering.
- **Price lookup:** Fetch normalized pricing for a specific provider or search pricing by model globally.
- **Cheapest provider lookup:** Instantly identify the most cost-effective host for a specific model across all supported providers.
- **Cost calculation:** Convert raw token usage into exact, itemized USD cost for any single request.

Canonical endpoints:

| Method | Path                                    | Purpose                                |
| ------ | --------------------------------------- | -------------------------------------- |
| `GET`  | `/v1/reference/providers`               | List supported providers               |
| `GET`  | `/v1/reference/providers/{provider_id}` | Get provider metadata                  |
| `GET`  | `/v1/reference/models`                  | List models                            |
| `GET`  | `/v1/price/{provider_id}`               | List model pricing for one provider    |
| `GET`  | `/v1/price/search`                      | Search pricing by model                |
| `GET`  | `/v1/find/cheapest/{model_id}`          | Find the cheapest provider for a model |
| `POST` | `/v1/calculate`                         | Calculate itemized request cost        |

Pricing values are normalized as USD per token. Our comprehensive pricing schema accurately captures edge-case dimensions, including cache reads, cache writes, reasoning tokens, web search, image, audio, and per-request fees.

Read the complete API reference at [docs.narev.ai/platform/api-reference/overview](https://docs.narev.ai/platform/api-reference/overview).

## 2. Why this matters

Narev Cloud is the foundational backend primitive for AI applications that demand exact, up-to-the-minute LLM pricing data.

Engineering and FinOps teams rely on Narev to:

- **Project AI unit economics** accurately before launching new features.
- **Calculate absolute Cost of Goods Sold (COGS)** for every single LLM request.
- **Run cross-provider comparisons** for the same underlying model.
- **Unlock cost arbitrage** by detecting when a lower-cost provider can serve the same workload.
- **Power usage-based billing** and precise margin analytics.
- **Maintain live internal model catalogs** without scraping provider pricing pages.

Instead of hardcoding prices or relying on stale, unmaintainable spreadsheets, developers query Narev to work with normalized, production-ready provider, model, and cost data.

### Trusted by the open source ecosystem

Narev is the authoritative pricing data source for the AI ecosystem. We already supply normalized pricing data to popular open source platforms, including:

- **models.dev** from Anomaly Co
- **Freebuff** from Codebuff
- **Mission Control** from Builderz.dev

## 3. The Narev Ecosystem

Narev Cloud is the hosted pricing and billing layer. It syncs live LLM pricing and exposes the Pricing API for applications that require current model cost data.

The **Narev SDK** is the runtime metering layer. It integrates seamlessly with the Vercel AI SDK and native provider SDKs to capture exact token usage, resolve costs via Narev Cloud, and route billable events to destinations like Stripe, Lago, Polar.sh, and OpenMeter.

**Narev Self-Hosted** is the open-source infrastructure cost layer. It empowers teams to map underlying infrastructure costs into the FOCUS format, enabling unified analysis of AI spend and traditional cloud COGS.

## 4. Benchmarking and optimization

Narev extends beyond pricing to support production-first model evaluation through the Benchmark Hub at [narev.ai/hub/benchmark](https://narev.ai/hub/benchmark).

The LLM-accessible benchmark list is available at [www.narev.ai/hub/benchmark/llms.md](https://www.narev.ai/hub/benchmark/llms.md).

Use benchmarks to rigorously compare cost, quality, and latency tailored to your specific tasks. Rely on the Pricing API when you need the underlying, live price data that fuels cost-aware routing, usage billing, procurement, or optimization workflows.

## 5. Integration context

Narev integrates natively into the tooling AI teams already use:

| Category                  | Integrations                                                   |
| ------------------------- | -------------------------------------------------------------- |
| Billing destinations      | Stripe, Lago, Polar.sh, OpenMeter                              |
| AI gateways and providers | OpenRouter, Portkey, LiteLLM, Helicone, AWS Bedrock, Vertex AI |
| Tracing and evaluation    | Langfuse, LangSmith, Weights & Biases                          |
| Agent tooling             | Narev agent skills and MCP server                              |

For agent workflows, install the Narev skills with `npx add skills narevai/skills`. Use the Narev MCP server when an assistant needs direct access to Narev pricing data and product context. Read the MCP setup guide at [www.narev.ai/docs/mcp](https://www.narev.ai/docs/mcp).

**The key idea:** Narev Cloud supplies the industry's most accurate LLM pricing data via API, while the SDK, agent skills, MCP server, and integrations transform that pricing data into precise billing, automated cost attribution, and crystal-clear unit economics.