# spaturzu

> spaturzu is an LLM cost attribution SDK. It attributes every OpenAI, Anthropic, Bedrock, Gemini, and Mistral API call to the AI agent, run, and project that caused it — in-process, with no proxy and no prompt storage.

The SDK meters each call inside your own process: it converts prompt and response text to token counts locally and sends only the counts and the computed cost. Your prompts, system prompts, messages, and the model's responses never reach spaturzu's servers — which suits teams with GDPR, HIPAA, or SOC 2 obligations.

## Guides

- [What is LLM cost attribution (and why your provider bill can't do it)](https://spaturzu.superchiu.org/blog/what-is-llm-cost-attribution): LLM cost attribution ties each API call's cost back to the agent, run, and feature that caused it. Here's what it is, why provider invoices can't tell you, and how to add it without a proxy.
- [How to track LLM API costs per agent in Node.js](https://spaturzu.superchiu.org/blog/track-llm-api-costs-per-agent-nodejs): A practical guide to attributing OpenAI and Anthropic spend to the agent and run that made each call in Node.js — with a drop-in SDK, no proxy, and a hard budget cap before the call.

## Product

- [Documentation](https://spaturzu.superchiu.org/docs): SDK setup for OpenAI, Anthropic, Bedrock, Gemini, and Mistral; per-agent attribution with run() and .withAgent(); in-process budget caps.
- [Pricing](https://spaturzu.superchiu.org/pricing): free to start, no card required.

## Comparisons

- [spaturzu vs Helicone](https://spaturzu.superchiu.org/compare/helicone): Helicone is an LLM observability platform with a full-featured AI Gateway — built to debug what a model actually said and to route requests across 100+ providers from one endpoint. spaturzu focuses on per-agent cost attribution and budget caps, in-process and without storing prompts.
- [spaturzu vs Langfuse](https://spaturzu.superchiu.org/compare/langfuse): Langfuse is an LLM engineering platform for tracing, prompt management, and evaluations — built to make a model's answers correct, with a first-class agent observation model. spaturzu focuses on per-agent cost attribution and budget caps, in-process and without storing prompts.

## Policies

- [Privacy Policy](https://spaturzu.superchiu.org/privacy): what spaturzu receives (token counts and cost metadata) and what it never receives (your prompts and responses).
- [Terms](https://spaturzu.superchiu.org/terms): beta terms of use.