Monitor OpenAI costsper feature.
Point your OpenAI client at Spillr. Tag each call with a feature name. See exactly how much email-summary, chat, or onboarding costs — not just a total bill.
Works with all GPT models · No prompt storage · 2-minute setup
import OpenAI from 'openai';
const client = new OpenAI({
baseURL: 'https://proxy.spillr.io/v1', // ← one line change
defaultHeaders: {
'X-Spillr-Key': process.env.SPILLR_KEY,
'X-Spillr-Feature': 'email-summary', // ← tag per feature
},
});
// The rest of your code stays exactly the same.
const res = await client.chat.completions.create({
model: 'gpt-4.1-nano',
messages: [{ role: 'user', content: prompt }],
stream: true, // streaming supported
});The problem
Your OpenAI bill is rising. Which feature caused it?
OpenAI shows you total daily token usage. It does not show you that your new AI-generated email digest is spending ten times more than your search feature. Without per-feature visibility, you are flying blind — and one runaway prompt can erase a month of margin.
No per-feature visibility
OpenAI's dashboard aggregates all usage. You can't see which feature is expensive.
No budget guardrails
There is no built-in way to block requests when a feature exceeds its spend limit.
Surprises at month end
Without real-time alerts, you only notice overspend when the invoice arrives.
Features
Everything you need to control OpenAI spend.
Cost per feature
Add one header to your OpenAI request. Spillr tags each call and groups costs by feature in the dashboard.
Budget alerts
Set a monthly or daily spend limit per feature. Get an email alert when you hit 80% — before you go over.
Hard stops
When a budget limit is reached, Spillr returns a 429 instead of forwarding the request. Your OpenAI bill stops at the limit.
Streaming support
OpenAI SSE streaming works unchanged through the proxy. Token counts are extracted from the final chunk.
Zero prompt storage
Spillr never stores your prompts or completions — only token counts and cost. Privacy by design.
Proxy or SDK
Use the proxy for hard stops and zero code changes, or the SDK wrapper for async reporting with full control.
FAQ
Common questions about OpenAI cost monitoring.
Does Spillr store my OpenAI prompts or responses?
No. Spillr only records token counts, the model name, and the calculated cost. Your prompt text and completion content are never stored — they pass straight through to OpenAI.
Which OpenAI models does Spillr support?
All of them. GPT-4.1, GPT-4o, GPT-4-turbo, GPT-3.5-turbo, o1, o3 — and any future model OpenAI releases. Spillr reads the model from each request and looks up the current token price automatically.
Does OpenAI streaming work through Spillr?
Yes. SSE streaming passes through the proxy unchanged with no added latency. Token counts are extracted from the final stream chunk, so streaming responses are tracked just as accurately as regular ones.
How is this different from OpenAI's built-in usage dashboard?
OpenAI's dashboard shows total token usage per day across your entire account. Spillr breaks costs down by feature label — so you can see that email-summary costs $14/month while your chat feature costs $2.
How much does Spillr cost?
Spillr is completely free during beta. Full feature access with unlimited projects, unlimited API calls tracked, and no credit card required.
Start monitoring OpenAI costs in 2 minutes.
Change one line. Add two headers. See per-feature costs immediately.
Create free account →