LLM Failover & Fallback API: Automatic Provider Redundancy
Automatic LLM failover across 220+ routes and 30+ providers. OpenAI-compatible: change base_url, keep your code. Requests roll over when a provider is down.
Your uptime should not depend on one provider's status page.
TrustedRouter is an OpenAI-compatible API with automatic LLM failover built in. When a provider goes down, requests roll over to a healthy route. No code change, no config push, no incident channel at 3 a.m. Your users keep getting answers.
One key fronts 220+ model routes across 30+ providers: Claude, GPT, Gemini, DeepSeek, Kimi, Qwen, GLM, Llama, Mistral, and more. Pin an exact model, or send traffic to trustedrouter/auto and let the gateway pick the best fit per request. Migration is a one-line change: point base_url at https://api.trustedrouter.com/v1 and keep your SDK, your model ids, and your code.
Browse routes See measured latency
New email signups receive starter credit. No card required.
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.trustedrouter.com/v1",
api_key=os.environ["TRUSTEDROUTER_API_KEY"],
)
resp = client.chat.completions.create(
model="trustedrouter/auto",
messages=[{"role": "user", "content": "Hello"}],
)
# Provider down? The request rolls over. Same code.Rollover is automatic
Outages are a fact of life on a single provider. The gateway handles fallback for you: when a provider is down, traffic moves to one that is up. There are no retry trees or fallback chains to maintain in your application.
Change base_url, keep the rest
The API is OpenAI-compatible. Swap the base URL and keep your SDK, your model ids, and your existing code.
Latency you can look up
A failover plan should rest on data. TrustedRouter publishes measured latency from real routed requests, per provider, so you can see how the alternate routes perform before an outage forces the question.
Failover should not mean spraying prompts across the internet.
Multi-provider redundancy has a cost: every fallback hop is another party that could see your data. TrustedRouter's prompt path runs inside a hardware-attested confidential-compute gateway. The source is open, the image digest is published, and the attestation binds to the live TLS session, so you can check it yourself. TrustedRouter never logs prompt or output content. Ordinary synchronous and streaming inference does not retain it. The opt-in Batch API temporarily retains enclave-encrypted artifacts for up to 30 days. Every route carries a privacy tier: Open, Zero-Data-Retention, or TEE, where even the serving provider cannot read the prompt. Attestation beats a privacy policy.
This holds at production scale. A document-processing company scaled from zero to over a billion tokens per week through the attested gateway within three weeks of onboarding. Start with prepaid credits, bring your own provider keys, or pay per request with crypto. Per-model rates are at /pricing. Try any route in the playground first.
Current routes, prices, privacy, and measured performance.
Catalog facts come from the routes currently configured in TrustedRouter. Performance uses the same cached metadata snapshot as the public leaderboard. Prompts and outputs are not part of these measurements.
| Model | Providers | Context | Input | Output | Privacy | Measured route |
|---|---|---|---|---|---|---|
Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 |
|
1,000,000 | $5.275/1M | $26.375/1M | varies 5 cited scores | 1353 ms TTFT anthropic · 90 tok/s · 100.00% available · n=2 |
OpenAI: GPT-5.5openai/gpt-5.5 |
|
1,050,000 | $5.275/1M | $31.65/1M | ZDR 3 cited scores | No measurements yet |
Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash |
|
1,048,576 | $1.5825/1M | $9.495/1M | ZDR | 1632 ms TTFT google-vertex · 72 tok/s · 100.00% available · n=2 |
MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code |
+11
|
262,144 | $0.692291/1M to $1.31875/1M | $3.4815/1M to $4.82135/1M | ZDR 5 cited scores | 1850 ms TTFT kimi · 23 tok/s · 100.00% available · n=21 |
Z.ai: GLM 5.2z-ai/glm-5.2 |
+28
|
1,048,576 | $0.3587/1M to $2.370374/1M | $1.5825/1M to $6.838973/1M | Confidential + ZDR 4 cited scores | 3220 ms TTFT phala · 68 tok/s · 85.71% available · n=60 |
MiniMax: MiniMax M3minimax/minimax-m3 |
+13
|
1,048,576 | $0.24265/1M to $0.633/1M | $1.0128/1M to $2.532/1M | ZDR 4 cited scores | 2234 ms TTFT minimax · 62 tok/s · 100.00% available · n=30 |
Browse every modelReview provider policiesOpen the full leaderboardSnapshot 2026-10-22T22:07:34.258Z
Questions
What is LLM failover and how does TrustedRouter handle it?
LLM failover means rerouting a request to another provider when the one serving your model goes down. TrustedRouter builds this into the gateway: automatic provider fallback rolls traffic over to a healthy route, drawing on 220+ model routes across 30+ providers. You can pin an exact model, or use the trustedrouter/auto meta-route and let the gateway pick the best fit per request.
Do I need to rewrite my code to get automatic provider fallback?
No. The API is OpenAI-compatible, so migration is a one-line change: point base_url at https://api.trustedrouter.com/v1 and keep your SDK, your model ids, and your existing code. Rollover happens inside the gateway, so there are no retry trees or fallback chains to maintain in your application.
Does routing through a failover gateway expose my prompts to more parties?
The prompt path runs inside a hardware-attested confidential-compute gateway. The source is open, the image digest is published, and the attestation binds to the live TLS session, so you can check it rather than take it on faith. TrustedRouter never logs prompt or output content. Ordinary synchronous and streaming inference does not retain it. The opt-in Batch API temporarily retains enclave-encrypted artifacts for up to 30 days. Each route carries a privacy tier: Open, Zero-Data-Retention, or TEE, where even the serving provider cannot read the prompt.
How do I know the fallback routes are fast enough for production?
TrustedRouter publishes measured latency data from real routed requests at /llm-provider-latency-benchmarks, so you can compare providers before an outage forces a switch. On scale: a document-processing company scaled from zero to over a billion tokens per week through the attested gateway within three weeks of onboarding.