OpenAI compatible API · Attested · Public status

Try TrustedRouter in 60 Seconds

Run one OpenAI-compatible request, inspect the live model catalog, and verify the attested gateway before moving real traffic.

Verify gateway

Onebase URL to migrate

100sof models and routes

Noneprompt logs by default

One requestOpenAI SDK compatible.

Live catalogModels, routes, and prices.

Fresh evidenceChallenge the gateway with your nonce.

Public statusCheck the service before a test.

60-second quickstart

Change the base URL. Keep the client.

Create an API key and export it as TRUSTEDROUTER_API_KEY.
Install the OpenAI client: pip install openai.
Run the request at right. A successful response should contain PONG.

Migration guide

QuickstartPython

import os
from openai import OpenAI

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

response = client.chat.completions.create(
    model="trustedrouter/zdr",
    messages=[{
        "role": "user",
        "content": "Reply with PONG only.",
    }],
    max_tokens=128,
)

print(response.choices[0].message.content)

Catalog

Inspect the live API.

Use the deployed model list as the source of truth for model IDs, context, capabilities, and pricing.

Open model JSON → Attestation

Verify before sending data.

Generate fresh nonce-bound evidence and compare the running image with the published open-source release.

Verify gateway → Operations

Check status and latency.

See router-core health separately from provider response health and regional measurements.

Open status →

Test it honestly

Four useful experiments.

Change only base_url in an existing OpenAI integration.
Run the same three prompts across an open model, a frontier model, and trustedrouter/cheap.
Use an ordered models list and leave provider fallback enabled.
Challenge the live gateway with your own nonce before the request.

State the boundary

What the proof does not claim.

TrustedRouter attestation verifies the hosted gateway workload. It does not prove that the source has no bugs, and it does not make every downstream provider end-to-end encrypted.

Prompt and output content is not durably stored by TrustedRouter by default. Provider retention, training, jurisdiction, and confidential-compute posture remain provider specific.

Provider posture Model routes Measured performance Source

Questions

Does TrustedRouter store prompts or outputs?

TrustedRouter does not durably store prompt or output content by default. Operational metadata includes model, provider, token counts, latency, cost, status, and region. Downstream provider handling remains provider specific and is published on model and provider pages.

What does gateway attestation prove?

A fresh nonce challenge lets you verify that the live prompt gateway is running the published workload image. It does not prove the code is bug free or make every downstream provider confidential.