AI observability + secure gateway

Protect AI keys.
Track LLM costs.

Track LLM costs, latency, errors and sessions across OpenAI, Claude, Gemini and more. Keep your existing backend, or add our AI gateway for API key protection, routing and fallbacks.

Start free Explore the demo

1,000 requests/month free · No credit card required · Provider usage billed separately

Request health at a glanceSample data
Request activity

Requests

SuccessError
1,233,710

98.79% success rate

AI spend$2,096.12Track spend
Avg. latency1.46sSpot slowdowns
Models4Compare models
TWO WAYS IN. ONE CLEAR PICTURE.

Your stack.
Your level of control.

Start with visibility. Add the gateway wherever you need protection and routing. Both can live in the same project.

Analytics only

Keep your routing.
See the whole picture.

Your backend calls the provider directly and sends completed request metrics to HyperProxy. Provider keys stay on your server.

  • Costs, tokens, latency and errors
  • Clients, sessions and quality annotations
  • No gateway service or provider key required
Connect analytics →

Server integration · metrics only · no prompt or response bodies

Gateway + analytics

Protect your keys.
Control your requests.

Route calls through HyperProxy. Keep provider credentials out of your app and collect request metrics automatically.

  • Split-key protection and optional App Attest
  • Provider-native routing, channels and fallbacks
  • Limits and versioned prompt releases
Add the gateway →

Mobile, web and backend apps · bring your own provider accounts

LESS GUESSWORK. MORE VISIBILITY.

Every spike tells a story.
See it in one place.

Requests, estimated AI spend, errors and latency — with shared periods and filters. Follow a model or client from the overview into the details.

HyperProxy · Project overviewInteractive preview · sample data
Workspace overview

Requests

SuccessError
1,233,710

98.79% success rate

Errors by status

429500502

Top models

ModelRequests
  1. GPT-4.1
    505.8K
  2. Claude Sonnet
    357.8K
  3. Gemini Flash
    234.4K
  4. GPT-4.1 mini
    135.7K

AI spend

$2,096.12

Estimated provider cost

Top clients

ClientRequests
  1. Production API
    555.2K
  2. iOS app
    345.4K
  3. Web app
    234.4K
  4. Development
    98.7K

Latency

1.46s / request

Average provider response time

Interactive product preview · illustrative data. Switch periods or hover over a chart to explore.
Connect my backend metricsSet up the gateway

Start with 1,000 requests per month. No card required; provider charges are separate.

Know where spend goes

Compare models and clients. Unknown pricing stays visible instead of looking free.

Find what changed

Inspect error codes and latency trends with the same date range across every chart.

Keep the context

Connect requests to clients and sessions. Separate gateway traffic from server-reported events.

SECURITY

Keep provider keys out of your app.

Two key halves unlock the provider credential in gateway memory for an authorized request. Neither half can recover it alone. Watch the path.

Protected request pathAutomatic replay
01

App inspected

Attacker finds only client_half.

No usable key
02

Database copied

Attacker gets ciphertext and server_half.

No usable key
03

Request in transit

The credential is recovered in memory while forwarding.

Provider-native

App Attest device attestation

Verify requests come from your real app on a real device. App Attest with Secure Enclave signatures, a replay guard (sign_count must strictly increase), and modes off / device_token / assertion — plus a rotatable simulator bypass token for dev. Play Integrity is on the roadmap.

INTEGRATION

Add the gateway when you need more control.

Point any request at HyperProxy and swap the auth header. That's it — the body stays your service's own.

request.sh
$ curl -N https://api.hyperproxyai.com/ab12cd34ef56aa77/c4bc1015a291/v1/chat/completions \
  -H "X-HyperProxy-Key: hp_live_ab12cd34ef56aa77.<key_id>.<client_half>" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-4o-mini","stream":true,"messages":[{"role":"user","content":"hi"}]}'

Copy your gateway URL from the dashboard and append the upstream's own path: api.hyperproxyai.com/<project>/<service>/<your upstream path>. The URL says where to route; the app key in the header is the credential — it carries the client half we fuse the real key from. Use the provider's request format. Configured model, prompt or compatibility policies may update fields before forwarding. Any public HTTPS API, AI or not.

REACH

Every major provider — or any API you add.

Start from a built-in preset, or register any public HTTPS API by its base URL — an AI model or any other service. Provider-native transit means no SDK lock-in: send each service's own request format and use native forwarding by default, with per-service endpoint allowlists and attestation.

OpenAI Anthropic Google Gemini xAI Grok DeepSeek Mistral Perplexity Novita Z.ai Stability AI Black Forest Labs Any custom API
iOS Android Web React Native Flutter
WHY HYPERPROXY

Visibility and control, from the same workspace.

Ships everywhere

Provider-native transit runs from iOS, Android, web, React Native, or Flutter — no SDK to adopt, no lock-in.

Defense in depth

Split-key envelope encryption plus App Attest device attestation — keep provider credentials off-device and add device verification to supported requests.

Cost you can see

Track tokens, model costs and spend limits. See when usage is unpriced so unknown costs do not look free.

Prompt releases without an app update

Test saved versions in staging, publish to production, and roll back from the dashboard. See the release workflow →

USE CASES

Start with the problem you need to solve.

Each guide explains the integration boundary, the data involved, and the controls available in production.

LLM cost tracking →

Measure requests, tokens, latency, errors and estimated provider spend—with or without the gateway.

AI gateway for Swift →

Call provider-native AI APIs from iOS without embedding the provider credential in your app.

Secure AI proxy →

Protect keys with split-key storage, device proof, allowlists, rate limits and model policy.

OpenTelemetry for LLMs →

Turn completed GenAI spans into focused usage and cost analytics over OTLP/HTTP.

CAPABILITIES

Observe, protect, and control your AI.

Split-key envelope encryption

client_half ⊕ server_half = data_key, reconstructed in memory only.

App Attest device attestation

off / device_token / assertion; Secure Enclave; replay guard; simulator bypass.

Named app keys

Many per service — segment and revoke per app version. Supply the provider key again when minting or rotating an app key.

Endpoint allowlists

Restrict which upstream paths a key can reach, per service.

Rate-limit rules

Enforce limits per key, IP, or device, with clear retry guidance for every rejected request.

Email alerts

Fire after N consecutive upstream failures — know before your users do.

First-class SSE streaming

Token-by-token streaming forwarded straight through, no buffering.

Self-host or managed

Run HyperProxy in your own environment, or let us operate it for you.

PRICING

Start free. Scale with production traffic.

Start with 1,000 requests per month, no card required. Choose by monthly request volume. Every plan includes the same observability and gateway controls, with usage-based pricing that scales with you.

Starter

Launch

For prototypes moving into production.

$9.99/ monthplus VAT / sales tax where applicable
10,000requests included / month
  • $1.50 / 1,000 requestsbeyond included usage
  • Account-wide usage controls
  • One allowance shared across the account
Choose Starter

Pro

Production

For production apps with steady traffic.

$49.99/ monthplus VAT / sales tax where applicable
100,000requests included / month
  • $0.75 / 1,000 requestsbeyond included usage
  • Account-wide usage controls
  • Lower unit cost for active products
Choose Pro

Premium

Scale

For growing products and multiple workloads.

$199.99/ monthplus VAT / sales tax where applicable
1,000,000requests included / month
  • $0.30 / 1,000 requestsbeyond included usage
  • Account-wide usage controls
  • Built for several production services
Choose Premium

Business

Volume

For high-volume production workloads.

$1,499.99/ monthplus VAT / sales tax where applicable
10,000,000requests included / month
  • $0.22 / 1,000 requestsbeyond included usage
  • Account-wide usage controls
  • Best unit economics at managed scale
Choose Business
Included with every managed plan

Full product access. Your plan sets included usage and usage-based pricing.

  • AI spend and model cost tracking
  • Request, error and latency analytics
  • Split-key gateway and security controls
  • Spend alerts, webhooks and prompt releases

Compare all plans, free allowance and billing examples →

Monthly prices are in USD and exclude VAT / sales tax, which is added based on the customer’s location. Provider API charges are separate. Every account includes 1,000 base requests per month; paid subscriptions add to that shared allowance. Usage-based rates apply after included usage, billed in blocks of 1,000 on the next invoice. Usage-based billing is on by default for paid plans and can be turned off in your account to stop at the included limit. When enabled, an account safety limit caps usage at 2× the total allowance.

DEPLOY

Run it yours, or let us run it.

SELF-HOST

Your infra, your keys.

Run HyperProxy in your own environment. Keep traffic, credentials, and operational control entirely on your side. Source access is available at no charge on request.

Ask about self-hosting
MANAGED

Skip the ops.

We run the platform. Connect server metrics or configure a secure gateway service. Free tier included, scale when you need it.

$ start free
GET STARTED

Start with visibility. Add control as you grow.

Ship AI features without shipping your API keys.

$ start free
$ start free Read the docs

Free tier: 1,000 requests / month. No card required.