Works with Cursor / Claude Code / Cline

Take control of
runaway AI agent costs.

Just change your base_url to add budget controls, auto-stop & failover.
Let your AI agents run overnight — without the bill shock.

Free plan is free forever — no credit card, sign up with Google

agent-monitor — daily limit: $1.00
$ agent.run("Build this app")
API calls: 0
Today's cost:$0.00/ $1.00 daily limit
qzira monitoring — agent running
Daily limits stop overnight runaway before it hits your wallet
OpenAI / Anthropic / Google AI supportedNo credit card requiredBYOK (use your own API keys)

Why you need a gateway

AI agents are powerful, but without controls they can rack up unexpected costs fast.

💬
One prompt
"Build this app"
🤖
Dozens to hundreds of API calls
The agent loops autonomously
💸
Unexpected costs
Bills can spiral without you noticing
10s–100s
API calls per single prompt
24h
Agents never sleep
Auto-stop
Blocks requests at limit (prevents unexpected costs)

Three shields to protect your wallet

qzira is not "just a proxy." It's a cost defense system for the AI agent era.

💰
Budget alert notificationsStarter+

Get email alerts when usage hits 50% and 80%. Catch cost spikes before they grow.

🛡
Daily limits + auto-stopPro+

Set daily caps on top of monthly limits. A kill switch that blocks requests at the limit — preventing overnight bill shock from runaway agents.

🔄
FailoverStarter+

If OpenAI goes down, auto-switch to Anthropic. Multi-provider redundancy that minimizes downtime risk.

Built for developers

Change one endpoint. Keep your existing code.

Multi-provider unified endpoint
OpenAI, Anthropic, Google — all through one endpoint. Prefix-based routing adapts to new models instantly.
BYOK (Bring Your Own Key)
Use your own API keys. No markup, no middleman margin. Just pay a flat monthly fee for gateway features.
Streaming support
Full SSE (Server-Sent Events) support. ChatGPT-like real-time responses across all providers.
Auto-retry
Automatic retries on transient errors. Exponential backoff to handle API rate limits gracefully.
Global edge network
Powered by Cloudflare Workers. Low-latency from anywhere in the world.
Usage dashboard
Visualize request counts, response times & error rates in real time. Make informed cost-optimization decisions.

Getting Started

Up and running in 3 easy steps

1

Sign up with Google

Done in 30 seconds. No credit card required.

2

Get your API key & register provider keys

Generate a qzira API key in the dashboard and register your OpenAI / Anthropic keys.

3

Change base_url to api.qzira.com/v1

Just update the endpoint in Cursor, Claude Code or Cline settings. No code changes needed.

Your dashboard after sign-up

See all your API usage, in one place

Request counts, token consumption & model-level logs in real time. Spot unexpected costs before they happen.

qzira dashboard — monitor API usage in real time

Real-time monitoring

Daily request counts & token usage displayed in charts

Request logs

View model, latency & status for every request

CSV export

Download usage data for detailed analysis

Compatibility

Works with your dev tools

OpenAI-compatible endpoint. Just change the URL in your settings.

Cursor

Verified on Pro and above

Claude Code

Messages API supported

Cline (VS Code)

Verified

Any OpenAI SDK-compatible tool

Just change base_url

One-line integration (SDK compatible)

Your existing OpenAI SDK code works as-is. Just change the baseURL.

app.py
from openai import OpenAI

client = OpenAI(
    api_key="gw_xxxxx",  # qzira API key
    base_url="https://api.qzira.com/v1"  # ← that's it
)

# business as usual
response = client.chat.completions.create(
    model="gpt-4o",       # or "claude-sonnet-4-20250514"
    messages=[{"role": "user", "content": "Hello!"}],
    stream=True
)

Affordable pricing (BYOK model)

Use your own API keys — no markup. Pay only for gateway features at a fair price.

Service
Lowest paid plan
Mid-tier plan
Portkey
$50
$200
Helicone
$20
$100
qzira
$5
$10

Why so affordable? qzira uses the BYOK (Bring Your Own Key) model. You pay AI providers directly, so there's zero markup from qzira. The monthly fee covers gateway features only (budget management, auto-stop, failover, etc.).

Security

Your API keys, kept safe

qzira is a BYOK (Bring Your Own Key) service. API key security is our top priority.

Encrypted storage

API keys are encrypted with AES-GCM. Never stored in plain text.

Processed on Cloudflare Workers

Requests are handled at the edge. Keys are decrypted only within the serverless environment — never sent externally.

BYOK — you own your keys

qzira holds no shared keys. You use your own API keys, so you can revoke or rotate them at any time.

Keys never logged

API keys are never included in usage logs. Even in the unlikely event of a log leak, your keys stay safe.

Simple, transparent pricing

Streaming included in every plan. Upgrade or downgrade anytime.

Free
$0/mo
1,000 requests
1,000 requests/mo
1 provider
1 API key
Streaming
Retry (1x)
Basic dashboard
Start free
Starter
$5/mo
10,000 requests
10,000 requests/mo
3 providers
2 API keys
Streaming
Retry (2x) + Failover
Budget alerts
Try free first
Business
$29/mo
500,000 requests
500,000 requests/mo
Unlimited providers
50 API keys
Streaming
Retry (3x) + Advanced failover
Budget alerts + Auto-stop
Response cache
Semantic cache
Priority support
Try free first
Scale
$69/mo
Enterprise requests
High-volume (custom limits)
Unlimited providers
Unlimited API keys
All features unlocked
Retry (5x) + Advanced failover
Budget alerts + Auto-stop
Semantic cache

Custom-tailored for reliable, long-term operation

Contact us

AI model usage fees are paid directly to each provider (BYOK model).

Limits may be adjusted for fair use.

Billed in JPY via Stripe. USD prices are approximate.

Coming soon

Unleash your AI agents
with confidence.

Try it now on the Free plan. Setup takes 30 seconds — just sign in with Google.

Start free in 30 seconds
No credit card30-second setupAPI keys encrypted