CloudRouter 文档

English

简体中文

English

简体中文

Appearance

CloudRouter FAQ ​

1. Product Questions ​

Q1: What is CloudRouter? ​

CloudRouter is a unified multi-model access platform that provides individual developers and enterprise teams with a unified, reliable, and cost-optimized solution for accessing large language models.

Q2: How is CloudRouter different from using the official API directly? ​

The main differences include:

  • Cost optimization: Save up to 80% on usage costs
  • Stability guarantees: Multi-channel redundancy, intelligent routing, and automatic failover
  • Unified management: Manage multiple models with a single account and a unified bill
  • Flexible billing: Offers both subscription and pay-as-you-go models

Q3: Which AI models are supported? ​

Currently supported:

  • Claude series: Opus 4.5, Sonnet 4.5, Haiku, and more
  • Continuously expanding to include more mainstream models

Q4: How do I choose the right plan? ​

  • Personal learning / light usage → Starter plan
  • Small team collaboration / frequent usage → Pro plan
  • Enterprise-grade applications / mission-critical workloads → Ultra plan
  • Flexible spending / uncertain usage → Pay-as-you-go
  • Special requirements / large-scale usage → Enterprise customization

Q5: Can I use a subscription plan and pay-as-you-go at the same time? ​

Yes. The two billing models do not conflict with each other, and you can flexibly choose between them based on different scenarios.

2. Technical Questions ​

Q6: How do I integrate with CloudRouter? ​

Get started in three steps:

  1. Contact customer support to obtain access
  2. Log in to the console (console.cloudrouter.online)
  3. Obtain an API Key and configure it in your application

Q7: How is data privacy and security guaranteed? Will my data be retained for training? ​

  • CloudRouter side: We do not retain users' request data ourselves; we only act as an API proxy to forward requests
  • Upstream model side: Whether Claude models retain data for training depends on the upstream provider's policies, which we cannot fully guarantee
  • Transmission security: HTTPS encrypted transmission is supported to ensure the security of the data transmission process
  • Enterprise solutions: The enterprise edition offers private deployment options, keeping your data fully within your control

Q8: Are the models you provide genuine Claude models? Could they be secretly swapped for other models? ​

  • Model authenticity: We provide a native Claude model pool, and all backend models are official Claude models (Opus/Sonnet/Haiku, etc.)
  • No swapping: We will never replace Claude models with other models such as GPT, GLM, or Kimi
  • Automatic switching mechanism: To ensure stability, we automatically switch within the Claude model pool when a model call fails, but we always guarantee a native Claude model
  • Transparency guarantee: We are committed to the authenticity and transparency of our model sources

Q9: What are the differences between the three channels (Value / Smart Optimized / High-Speed Stable)? ​

  • Value Experience Channel (¥0.6/$): Lowest cost, suitable for testing, validation, and non-critical applications, with baseline stability and speed
  • Smart Optimized Channel (¥1/$): Balances cost and performance, suitable for daily development and general business, with good stability and fast response
  • High-Speed Stable Channel (¥5/$): Highest cost, suitable for production environments and mission-critical workloads, with optimal stability and the fastest response
  • Channel mechanism: Different channels correspond to different levels of our investment in cost, stability guarantees, and concurrency capacity. The more expensive the channel, the greater the investment in stability, speed, and concurrency

Q10: How is stability guaranteed? ​

  • Multi-channel redundant architecture
  • Intelligent routing with automatic failover
  • Intelligent scheduling across multiple providers
  • Enterprise-grade SLA guarantees (Ultra plan and above)

3. Business Questions ​

Q11: How do I pay? ​

  • Subscription plans: Support Alipay, WeChat Pay, and corporate bank transfer
  • Pay-as-you-go: Charged based on actual usage after topping up
  • Enterprise customization: Supports corporate contracts and invoices

Q12: What is the refund policy? ​

  • Starter plan: 7-day no-questions-asked refund
  • Pro plan: 30-day no-questions-asked refund
  • Ultra plan: Flexible negotiated refund mechanism
  • Refund amount is calculated based on the remaining unused days

Q13: Can I upgrade or downgrade my plan? ​

Yes. Plan changes take effect in the next billing cycle.

Q14: How are invoices issued? ​

  • Individual users: Electronic invoices can be issued
  • Enterprise users: Special VAT invoices are supported
  • Contact customer support to provide invoicing information

Q15: Is there a trial period? ​

  • New users can contact customer support to apply for a trial quota
  • Subscription plans come with a refund guarantee, so you can try them with confidence

Document version: v1.0 (2026-01-30)

Contact: support_cloudrouter@clouditera.com