Own your alpha

Every model.Provable privacy.

One OpenAI-compatible API for hundreds of open-weight and frontier models. Requests cross an attested gateway that does not log prompt or output content. Even our engineers cannot read your requests.

Verify trust

Migrate from OpenRouter OpenAI-compatible API Latency benchmarks

End-to-End Encrypted AI gateway ZDR routing EU routing 100% open source

3Live regions

HundredsModels through one API

NeverPrompt logs

PublicSource and status

Why developers choose TrustedRouter.

The trust model is the product.

TrustedRouter is an OpenRouter alternative for teams that need verifiable privacy, lower-cost model choices, provider fallback, and one API across hundreds of models.

Better trust

Match the hosted gateway to a public source commit and image digest before production traffic moves.

Cheaper routes

Use Qwen, GLM, DeepSeek, Google Gemma, Kimi, MiniMax, and other open models beside every frontier model.

Faster migration

Keep the OpenAI SDK and change one base URL. Agents can read llms.txt or use the MCP server.

More reliable inference

Provider failover and regional routing keep one upstream outage from becoming your product outage.

One compatible endpoint

Move over with one base_url change.

Keep your existing OpenAI-shaped calls. Add privacy and routing constraints only when you need them.

Use trustedrouter/auto for healthy-provider rollover
Use trustedrouter/zdr for zero-retention routing
Use trustedrouter/e2e for confidential + E2EE provider routes
Use trustedrouter/eu for Europe-focused routing
Bring your own provider keys where needed

Read the migration guide

OpenAI SDKPython

from openai import OpenAI

client = OpenAI(
    api_key=os.environ["TRUSTEDROUTER_API_KEY"],
    base_url="https://api.allyrouter.com/v1",
)

response = client.chat.completions.create(
    model="trustedrouter/auto",
    messages=[{"role": "user", "content": "Hello"}],
    extra_body={
        "provider": {"min_privacy": "confidential"}
    },
)

One API · hundreds of models

Choose a model, or choose a routing promise.

Named aliases make privacy, geography, cost, and failover explicit. Direct model IDs remain available when exact model selection matters.

Model	Designed for	Routing policy
`trustedrouter/auto`	General production traffic	Healthy provider rollover
`trustedrouter/zdr`	Legal and sensitive work	Zero-retention providers
`trustedrouter/e2e`	Confidential workloads	Confidential provider compute
`trustedrouter/eu`	European workloads	EU-focused routes
`trustedrouter/synth`	Combination models	Multiple models, one answer

Browse all models Compare providers

Verify trust

Do not take our word for any of this.

The production prompt path is separate from the dashboard and billing control plane. Every trust claim reduces to evidence you can inspect.

Your applicationOpenAI-compatible request

Encrypted transport ↓

Attested TrustedRouter gatewayPrompt path · no content logs

Selected route ↓

Model providerProvider policy published separately

Verification path Public

01read the source

02check attestation

03match digest to commit

04verify fail closed

Open live trust evidence →

Regional reliability

Routing that does not depend on one provider or one region.

Regional gateways, provider health measurements, and explicit fallback keep failures visible and contained. The public status page separates router-core health from upstream model health.

3 live API regions
Provider and model health measured continuously
Public latency and availability history
No silent fallback to a non-attested prompt path

View public status

10 regions across three clouds 3 live GCP regions, plus standalone EU (AWS) and APAC (Azure) deployments. Additional capacity staged.

100% hosted open source

A router you can inspect, fork, or run yourself.

Frontend, backend, gateway, infrastructure, SDKs, and deployment configuration are public. Hosted attestation connects that source to the code serving requests.

quill-routerControl plane and website trusted-router-pyPython SDK trusted-router-jsTypeScript SDK quill-cloud-infraDeployment infrastructure

No subscription required

Pay for what you route.

Use prepaid credits, BYOK, Stripe, PayPal, or stablecoin. Provider prices and routing costs remain visible before you call.

Pay as you go
Bring your own provider key
Automatic refill with a saved card
Migration credits by approval

Review pricing

Questions

What developers ask before routing.

What is TrustedRouter?

A privacy-first AI gateway that routes requests across many providers through one OpenAI-compatible API without logging prompt or output content.

Is this an OpenRouter alternative?

Yes, with an attested prompt path, public source, explicit privacy aliases, and fail-closed behavior as core design constraints.

Do you log prompts or outputs?

No. Prompt and output content are intentionally excluded from the control plane and metadata systems. Optional content export requires explicit destination-level configuration.

What does attested gateway mean?

The gateway runs in a Trusted Execution Environment and publishes evidence that connects the running image to public source. You can verify that evidence on the trust page.

Can I bring my own provider keys?

Yes. BYOK lets you retain provider commitments or enterprise rate limits while using the TrustedRouter routing and trust boundary.

What happens if attestation fails?

The gateway fails closed. It does not silently route prompts through an unverified fallback.

How do I migrate?

Keep the OpenAI SDK, change the base URL, and use the same chat or Responses calls. The migration guide includes tested examples.

Start with evidence

Every model. One verifiable prompt path.

Verify trust Migrate