LLM cost trackingper feature.
Ship AI features without flying blind. Spillr tracks every LLM API call — tagged by feature — across OpenAI, Anthropic, and Gemini from one dashboard.
OpenAI · Anthropic · Gemini · Proxy or SDK · 2-minute setup
import OpenAI from 'openai';
// Works the same for OpenAI, Anthropic, and Gemini —
// change baseURL per provider docs.
const client = new OpenAI({
baseURL: 'https://proxy.spillr.io/v1',
defaultHeaders: {
'X-Spillr-Key': process.env.SPILLR_KEY,
'X-Spillr-Feature': 'search', // ← your feature label
},
});
const res = await client.chat.completions.create({
model: 'gpt-4.1-nano',
messages: [{ role: 'user', content: prompt }],
});The problem
Your LLM bill is a black box until month end.
Solo devs shipping multiple AI features get one invoice from each provider — with no breakdown by product surface. One runaway loop or expensive model swap can erase your margin before you notice.
Invoice-only visibility
You see the total after the fact. Not which feature, user flow, or model drove the spike.
Multi-provider sprawl
OpenAI for chat, Anthropic for docs, Gemini for search — three bills, zero unified view.
No automatic limits
Without guardrails, a bug can burn unlimited tokens until someone checks the dashboard.
Features
Track LLM spend the way you ship features.
Feature labels
Tag every call with X-Spillr-Feature. Costs group by label in real time — search, chat, OCR, whatever you ship.
Live cost dashboard
See USD spend per feature, per model, and per provider. Updated on every API call.
Budget alerts
Set monthly or daily limits. Email when you hit 80% — before the invoice surprises you.
Hard spend stops
Proxy mode returns 429 when a budget is exhausted. Your LLM bill stops at the cap.
Privacy by design
Token counts and cost only. Prompts and completions are never stored.
Drop-in integration
Change base URL + two headers, or wrap with track(). No rewrite required.
FAQ
Common questions about LLM cost tracking.
What is LLM cost tracking?
LLM cost tracking means measuring how much you spend on large language model API calls — broken down by feature, model, or provider — instead of relying on a single monthly invoice total.
Which LLM providers does Spillr track?
OpenAI, Anthropic, and Google Gemini out of the box. All costs appear in one dashboard with per-feature breakdowns and cross-provider budgets.
Do I need to change my application architecture?
No. Point your existing SDK client at the Spillr proxy, or wrap calls with the Spillr SDK. Your request/response code stays the same.
Does Spillr store prompts or completions?
No. Spillr records only token counts, model names, feature labels, and calculated USD cost. Prompt content never touches our database.
How is Spillr different from provider usage dashboards?
Provider dashboards show totals per API key. Spillr adds the feature dimension — so you know which product surface is expensive, not just that your bill went up.
Start tracking LLM costs in 2 minutes.
One proxy URL. Two headers. Per-feature visibility on your next deploy.
Create free account →