Legal Terms & Privacy Policy
Last updated: June 22, 2026
01/Service Overview
AnnexAPI (the "Service") acts as a high-speed, unified intelligent API proxy routing gateway. We aggregate access to multiple upstream foundation large language models (including OpenAI, Anthropic, Google Gemini, Zhipu AI, and others) through a single, standardized OpenAI-compatible API interface.
By registering an account and calling our custom endpoints, you contract directly with us for proxy transit and load balancing. You acknowledge that we do not generate the AI content ourselves and merely act as a conduit to upstream API servers.
02/Account & Key Security
You are solely responsible for maintaining the confidentiality of your account credentials and any secret API keys issued to you. All request billing, traffic routing, and generation activities originating from your API credentials will be deemed your authorized actions.
You agree to notify AnnexAPI immediately if you suspect unauthorized key exposure. We reserve the right to cycle, revoke, or suspend any API keys exhibiting anomalous request patterns.
03/Billing & Credit Deductions
Our service operates under a volume-based pay-as-you-go credit structure. Charges are calculated dynamically on a per-request basis according to the usage metrics logged by our service provider.
- Any purchased credits are non-refundable and all deposits/top-ups are final.
- Credit deductions are calculated directly from downstream service logs (using the formula: logged quota divided by 500,000 to determine Aurea spent).
- Due to asynchronous background synchronization checks, there may be a slight latency before the final credit deduction is reflected on your dashboard.
- Costs logged on your dashboard are approximate until fully synchronized from the service provider logs.
04/Compliance & Content Moderation
AnnexAPI requires all customers to implement strict content filtering mechanisms on their customer-facing frontends. You must not call our API endpoints to submit or generate content that:
- Violates any domestic or international laws, regulations, or policies.
- Generates politically sensitive, defamatory, extreme, or hate-speech related materials.
- Bypasses security, safety, or alignment restrictions of the upstream models (such as jailbreaking or prompt injection attempts).
- Aids in cyber-attacks, reverse engineering models, or illegal network exploration.
You agree to comply with the respective model provider Terms of Service (such as the OpenAI Usage Policies and Anthropic AUP) when sending requests to those engines through AnnexAPI.
05/Upstream Dependencies & Disclaimers
Because AnnexAPI functions as an intelligent middleware router, our service availability, generation speed, and response quality depend entirely on the operational status and upstream network paths of third-party model providers.
We do not guarantee that the Service will be 100% uninterrupted, error-free, or meet any specific real-time latency thresholds under all conditions. All services are provided on an "as is" and "as available" basis without warranties of any kind, whether express or implied.
06/Account Suspension & Forfeiture
We reserve the right to temporarily suspend or permanently terminate your account, cycle your API access, or block query routing at our sole discretion, without prior notice, in the event of:
- Breach of any compliance or acceptable use directives.
- Repeatedly generating content flagged as unsafe or illegal by upstream providers.
- Abusive requesting patterns (such as deliberate DDoS-like requests or attempting API parameter exploration).
Upon permanent account termination due to safety or legal violations, all unused credits, balances, or outstanding top-ups associated with your profile will be immediately and irrevocably forfeited.
07/Privacy Policy
AnnexAPI respects your privacy. This Privacy Policy details how we handle the information that passes through our transit API gateway.
7.1 Information We Collect
We collect profile credentials (such as name, email, hashed password, and transit API keys) to enable secure system authentication and load-balancing services. We collect usage telemetry data (including timestamp, model, latency, response outcomes, and tokens consumed) for system billing, rate-limiting, and audit verification.
7.2 Data Transmission Security
We do not inspect or store the contents of prompt requests or completion response payloads. All requests are securely proxied through encrypted HTTPS tunnels directly to model providers.
7.3 Cookies & Verification
We utilize secure HTTP-only session cookies and OTP systems to prevent session hijacking and verify account profile modifications.