Introduction

CostCanary tracks per-user LLM costs in real time. Know exactly which users, features, and models are driving your AI bill.

What is CostCanary?

CostCanary is an SDK + dashboard that gives you per-user, per-feature visibility into your LLM spending. No proxies, no vendor lock-in — just one line of code around your existing LLM calls.

Why it exists

Most teams fly blind. They see a monthly OpenAI bill but have no idea which users, features, or models are driving it. CostCanary answers three questions:

Who is using AI? — Per-user cost attribution
What are they using it for? — Per-feature breakdown
How much does it cost? — Real-time cost tracking across all providers

How it works

javascript

Your app → CostCanary SDK → Track LLM calls → Dashboard
                                        ↓
                                Per-user costs, feature breakdown, model usage

1. Install the SDK: `npm install @costcanary/sdk`

2. Wrap your LLM calls with `track()`

3. Watch costs appear in your dashboard within seconds

Supported providers

OpenAI, Anthropic, Google Gemini, Groq, Mistral, Cohere, Perplexity, Together AI, Fireworks AI, Azure OpenAI, OpenRouter, Replicate, Anyscale, and any HTTP-based LLM API via auto-capture.