For startups shipping AI features

Ship AI without
eating the bill.

Fund a budget pool. Allocate spend per user. Set hard caps that stop calls before they overspend. Zorveus handles the meter, the billing, and the dashboards — so you ship product, not infrastructure.

Start free

No provider accounts to manage · Bring your own key supported

The problem

AI features ship fast.
The billing problem ships with them.

Unpredictable provider costs

One power user runs 10,000 summaries and your OpenAI bill triples. No warning, no cap, no way to attribute it.

Building billing instead of product

Wallets, metering, caps, webhooks, usage dashboards. Six weeks of engineering on something nobody wants to own.

No per-user visibility

You know the total bill. You don't know which customer drove it, which model they hit, or what it cost you vs. them.

Can't charge users fairly

Free tier eats your margin. Power users want more. You have no clean way to allocate, cap, or let them top up.

How it works

Four steps. No billing infrastructure.

You fund the pool and set the rules. Zorveus sits between your product and the model providers, checking every request before it costs you money.

01

Fund a budget pool

Top up your company wallet once. One balance funds every product user's AI usage. USD or NGN.

02

Allocate per user

Give each user a monthly AI allowance. Power users can top up their own extra spend if you allow it.

03

Set hard caps

Per-user, per-app caps. We check every request against the cap before the provider is called. No overspend.

04

Ship your product

Your backend calls Zorveus with a zrv inference key. We handle metering, billing, dashboards, and attribution.

The dashboard

See every user.
Control every dollar.

Budget pool, per-user burn, model breakdown. The dashboard tells you exactly who is spending what — and stops them before they overspend.

Zorveus
OA

Budget pool

Orceum Inc. · USD

Available in pool

$250.00

of $330.21 funded this month

Month spend

$80.21

5,100 requests

Active users

1,200

140 new this month

Pool health

Healthy

At current burn rate, pool lasts ~12 more days. Top up to cover the month.

Zorveus
OA

Product users

Per-user AI burn · this month

1,200 active
UserSpend / CapPlan
A

Ada

1.2M tokens

$12.40

pro
K

Kofi

82K tokens

$0.80

free
M

Mira

2.1M tokens

$18.20

pro
J

Jude

31K tokens

$0.30

free
View all 1,200 users

The difference

Your users spend.
You set the ceiling.
We handle the meter.

Every AI request from your product flows through Zorveus. We attribute it to the right user, check their cap, route it to the provider, and bill your pool — all in real time, before the cost is incurred.

Per-user attribution

Know who spent what

Hard caps

Stop calls before overspend

Bring your own key

Use your negotiated pricing

Integration

One key.
One gateway.

A single inference key is all you need. Point your OpenAI SDK at Zorveus, pass the user in metadata, and we handle attribution, caps, and billing — all in real time.

  • One inference key (zrv_...) — same OpenAI SDK you already use
  • Pass external_user_id in metadata for per-user attribution
  • Product user details auto-created on first call — no separate signup
  • Bring your own provider key and keep your negotiated pricing
app.py
from openai import OpenAI

client = OpenAI(
    base_url="https://api.zorveus.com/v1",
    api_key=zrv_inference_key,
)

response = client.chat.completions.create(
    model="openai/gpt-4.1-mini",
    messages=[{"role": "user", "content": "Hi"}],
    extra_body={
        "metadata": {
            "external_user_id": "cus_12345",
            "product_user": {
                "display_name": "Ada Lovelace",
                "email": "ada@example.com",
                "metadata": {
                    "plan": "pro",
                    "workspace_id": "workspace_789",
                    "source": "my_app"
                }
            }
        }
    },
)

# Zorveus attributes the cost to Ada,
# checks her cap, and bills your pool.

Stop building billing.
Start shipping AI.

Set up a budget pool in minutes. Attribute every call. Cap every user. Let Zorveus handle the rest.

Start free