CostRouter FAQ | 10 Questions We Hear Most Often

Lower costs shouldn't mean compromising on reliability, model quality, or security. Here's how CostRouter works.

Q1. Why are your model APIs so affordable?

We operate our own account resource pool（Including official Azure and Microsoft channels）and optimize sourcing across multiple channels. Our monthly credit-card spending exceeds US$200,000. Scale and efficient routing allow us to offer competitive pricing without deliberately downgrading models.

 

Q2. Is CostRouter reliable for production?

Yes. Our gateway is built for production applications and business workflows, with load balancing and latency-aware routing. If a route slows down or fails, eligible traffic can switch to another compatible route within seconds, when available.

Q3. Can you issue invoices and accept corporate payments?

Yes. We support invoices, B2B agreements, and corporate bank transfers.

Q4. Why is online card payment unavailable?

After a large wave of fraudulent card payments, we've temporarily disabled online card checkout. USDT remains available, while B2B payments and recharges can be arranged manually.

Please contact sales@costrouter.ai or support@costrouter.ai.

Q5. Are there speed, performance, or usage restrictions?

We provide the requested model with its native capabilities and safety boundaries. Provider policies still apply. Throughput, concurrency, and latency can differ by route; enterprise capacity can be planned with our team.

Q6. Do you retain prompts or model outputs?

For our enterprise service, we do not retain prompts, uploaded files, or model outputs. We keep usage records needed for billing and reconciliation. Relevant upstream provider policies still apply.

Q7. How can I protect my API keys?

You can set spending limits, expiration dates, model restrictions, and IP allowlists. We recommend separate keys for each application and environment.

Q8. Do lower prices mean weaker safety controls?

No. CostRouter does not remove the underlying model's safety policies or acceptable-use restrictions.

Q9. What happens if a provider route goes down?

Where an alternative route for the same requested model is available, our gateway can fail over without silently switching you to a different model. Provider-wide outages may still affect service.

Q10. Can you support high-volume enterprise workloads?

Yes. We offer tailored routing, concurrency planning, enterprise billing, and B2B contracting.