title: FAQ description: Frequently asked questions about AIOrouter — billing, security, models, and more.

Frequently Asked Questions

General

What is AIOrouter? AIOrouter is a Canadian AI API proxy that gives BETA customers access to approved DeepSeek, Qwen, GLM, Kimi, Gemini, and Claude models through a single OpenAI-compatible API. It uses monthly subscription billing with a provider-cost indexed allowance, Canada-resident gateway infrastructure, and built-in privacy protection controls — including bidirectional PII pseudonymization, technical secret redaction, and an AI Firewall.

Who is AIOrouter for? Canadian developers and enterprises who want privacy protection when using AI models. Also great for any developer who wants predictable monthly AI costs.

Pricing & Billing

How much does it cost? AIOrouter is pay-as-you-go: one-time Top-Up packs (Starter $10, Builder $50, Scaler $200, Pro $500, CAD) credit your balance, and each API call deducts from it at the public per-token rates in the Model Catalog. The checkout total is itemized before you pay (balance credit always equals the pack face value).

Is there a free tier? New accounts get a one-time free trial — 25,000 tokens over 7 days, model-locked to deepseek-v4-flash. After the trial, purchase a Top-Up pack to continue — your same API key keeps working. Trial users may also Top-Up during the trial; the first successful Top-Up permanently ends the trial.

How does usage-based billing work? Every call is estimated against your balance before dispatch; 402 Payment Required (balance_required, top_up) is returned when the balance cannot cover the request, so spend never goes negative. Actual usage is deducted after the call at the provider's public cache-aware token rates converted to CAD with the daily FX rate.

Can I cancel anytime? There is nothing to cancel — Top-Up is pay-as-you-go with no recurring billing. Your balance never expires; use it whenever you like.

Security & Privacy

Where is my data processed? Primary gateway processing happens in Montreal, Canada. Requests pass through privacy filtering before protected content is forwarded to AI providers.

How does AIOrouter protect my data? AIOrouter provides layered privacy protection: bidirectional PII pseudonymization (detects and replaces personal data with template tags before routing to providers), technical secret redaction (API keys, passwords, tokens), and an AI Firewall (prompt injection and jailbreak detection). Privacy protection is configurable per API key via the Dashboard. All processing runs on Canada-resident infrastructure (GCP Montréal). See our Privacy Policy for details.

Can I disable privacy protection? Yes. You can configure an individual API key to bypass privacy filtering via the Dashboard "Disable Shield" toggle. When Shield is disabled, your raw prompts are sent directly to AI providers without filtering — you accept full responsibility for any data transmitted. See our Terms of Service §5.4 for details.

Do you store my prompts? API prompts are processed in memory and not persistently stored. Audit logs record metadata (timestamp, model, token count) but not prompt content. See our Privacy Policy for the full data flow.

Do the Chinese AI providers see my data? Yes — protected request content is forwarded to the selected AI provider (e.g., DeepSeek) after privacy filtering. This is inherent to API proxying. Review each provider's terms for their data handling policies.

Technical

Is the API OpenAI-compatible? Yes. Drop-in replacement — change base_url and api_key, keep everything else.

Which models are available? DeepSeek V4 Pro, DeepSeek V4 Flash, Qwen 3.7 Max, Qwen 3.6 Flash, Qwen 3.6 Plus, Qwen3-235B, GLM-5.1, Kimi K2.6, Gemini 3.1 Pro Preview, Gemini 2.5 Pro, Gemini 2.5 Flash, and Gemini 3.5 Flash Preview are listed for BETA review. See Model Catalog.

Does AIOrouter support streaming? Yes. SSE streaming works exactly like OpenAI's — set stream: true.

Does AIOrouter support function calling? Yes. Most models support tool/function calling. See Model Catalog for per-model capabilities.

What happens if a provider is down? AIOrouter's intelligent router automatically falls back to an alternative provider. You can see which provider handled your request in the X-Provider response header.

What's the latency? ~15-50ms to Canadian developers (GCP Montreal). Provider response times vary by model and load.

Can I use multiple models with one API key? Yes. One API key gives BETA accounts access to the approved model catalog. Switch models by changing the model parameter.

Account & Support

How do I get an API key? Create an account at aiorouter.ca, verify your email, enroll MFA, add Top-Up credits (or start the free trial), and then generate your API key.

What if I lose my API key? You can generate new keys from the Dashboard. Old keys can be revoked for security.

How do I contact support? Email support@aiorouter.ca. Response times: P2 within 24h, P1 within 4h, P0 within 15min.