Terms of Service

By accessing or using the Infyrence platform, APIs, dashboard, and related infrastructure services (collectively, the "Services"), you agree to be bound by these Terms of Service. If you are using the Services on behalf of an organization, you represent and warrant that you have the authority to bind that organization to these Terms.

If you do not agree to these Terms, you may not access or use the Services.

Infyrence provides an edge-native AI gateway, proxy, and optimization router built on Cloudflare Workers (V8 isolates) and the Hono framework. The Services enable developers and organizations to access multiple large language model providers through a single, unified OpenAI-compatible API endpoint. Core capabilities include:

-Inference Routing: Latency-aware, multi-provider request routing with automated failover chains, exponential backoff, and jitter to eliminate cascading failures.-Virtual Key Management: Cryptographic Virtual Key generation, distribution, and lifecycle management scoped to organizations.-Usage Dashboard: Real-time telemetry including token consumption, latency percentiles, cost projections, and error rate analytics.-Hierarchical Rate Limiting: Enforced RPM and TPM quotas across organization keys, individual model tiers, and global infrastructure boundaries.-Edge-Level SSRF Shield: Active protection against Server-Side Request Forgery attacks targeting internal or unauthorized network topologies.

We reserve the right to modify, suspend, or discontinue any part of the Services at any time with reasonable notice.

Access to the Infyrence gateway is controlled through cryptographic Virtual Keys generated under your organization's account metadata. By generating Virtual Keys, you accept the following terms:

Sole Responsibility: You bear absolute responsibility for safeguarding, distributing, rotating, and monitoring all Virtual Keys issued under your account. This includes restricting key access to authorized personnel and enforcing key-scoped permissions where applicable.

Compromised Credentials: Infyrence cannot be held liable for credit exhaustion, unauthorized usage, or data exposure resulting from compromised, leaked, or improperly distributed client credentials. You are responsible for immediately revoking any key suspected of compromise through the organization dashboard.

Key Revocation: You may revoke any active Virtual Key at any time. Revoked keys are immediately invalidated across all edge nodes without propagation delay. In-flight requests using a revoked key will receive a 401 Unauthorized response.

No Liability for Credential Misuse: Infyrence processes every request bearing a valid Virtual Key signature. We are not obligated to verify the identity of the human or system presenting a key - the key itself constitutes authorization.

The Infyrence gateway enforces strict hierarchical rate limiting across three isolation scopes to ensure fair usage and infrastructure stability:

Organization / API Key Level: Each Virtual Key is assigned per-minute request (RPM) and per-minute token (TPM) quotas defined by your organization's subscription tier. Exceeding these quotas results in HTTP 429 (Too Many Requests) responses with a Retry-After header.

Model Tier Level: Individual model endpoints (e.g., gemini-2.0-flash vs. gemini-1.5-pro) carry independent rate limits enforced at the edge. High-throughput models may have lower per-key ceilings to preserve equitable access across the platform.

Global Infrastructure Limits: Platform-wide RPM and TPM ceilings protect the shared infrastructure from saturation. These limits apply regardless of individual key quotas and cannot be increased through subscription upgrades.

Circumvention and Abuse: Any intentional attempt to circumvent, saturate, bypass, or execute Distributed Denial of Service (DDoS) attacks against these rate-limiting mechanisms will result in immediate and automated programmatic blockages of the offending organization. This includes but is not limited to: key rotation to reset counters, distributed key usage to aggregate tokens, or payload fragmentation to evade token-level limits. Blocked organizations forfeit access to the Services without prior notice.

Model Pricing: Inference costs are passed through at provider list pricing. Pricing may change with 30 days' notice posted to the platform changelog.

Complimentary Credit Program: New developer organizations receive 50 complimentary credits upon registration (1 credit = 1,000 tokens). These credits are non-transferable, possess no monetary value, and cannot be redeemed for cash. Infyrence reserves the right to throttle, modify, or revoke complimentary credits if a team's usage patterns indicate automated script spamming, credential sharing, or abuse of the free tier.

Subscription Plans: Paid subscription fees are billed in advance on a monthly or annual basis via Stripe. All fees are non-refundable except as required by applicable law.

Late Payments: Overdue amounts may accrue interest at 1.5% per month. We may suspend your API access for non-payment after 15 days of delinquency.

You agree not to use the Services for the following prohibited activities. Violations result in immediate account termination and forfeiture of remaining credits and data:

Server-Side Request Forgery (SSRF): You may not utilize the gateway to perform SSRF attacks, attempt to scan internal IP blocks (127.0.0.0/8, 10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16), probe loopback networks, or target unauthorized cloud private endpoints through the inference routing pipeline.

Reverse Engineering and Tampering: You may not attempt to reverse-engineer the underlying routing pipeline, intercept upstream provider responses outside the authorized API contract, or tamper with the non-blocking logging workflows managed via edge runtime execution contexts (`c.executionCtx.waitUntil`).

Provider Spoofing: You may not modify custom extra headers or payload structures to deliberately spoof, impersonate, or manipulate upstream AI infrastructure providers, or misrepresent request origin to circumvent provider-level rate limits or content policies.

Credential Attacks: You may not attempt to brute-force, dictionary-attack, or otherwise exploit the Virtual Key authentication system, or test unauthorized keys against the gateway.

Infrastructure Disruption: You may not execute traffic amplification, request flooding, or any pattern designed to degrade the availability or performance of the edge infrastructure for other tenants.

Our IP: The Services, including the routing algorithms, edge infrastructure code, platform software, dashboard interfaces, documentation, and branding, are owned by Infyrence and protected by applicable intellectual property laws.

Your Content: You retain all rights, title, and interest in your prompts, inputs, and model completions. Infyrence claims no ownership, license, or training rights over your content. Your prompts and responses are your intellectual property.

Feedback: If you provide feedback or suggestions about the Services, you grant us a non-exclusive, royalty-free license to use that feedback for any purpose.

Model Providers: Each upstream model provider retains their own intellectual property rights. Use of specific models is subject to the provider's terms of service and acceptable use policy.

Service Availability: Infyrence utilizes an advanced multi-provider failover engine with automated exponential backoff and jitter to achieve a 99.9% uptime target across the gateway infrastructure. However, the Services are provided "as is" and "as available" without guarantees of uninterrupted availability.

No Warranty: We do not warrant that the Services will be error-free, latency-free, or secure, or that all defects will be corrected. We are not responsible for the accuracy, reliability, availability, or content policy compliance of any third-party model accessed through our platform.

Latency Metrics: Latency metrics displayed on the platform represent historical rolling averages and are not binding SLA commitments. Actual request latency varies based on provider load, model complexity, and network conditions.

Scheduled Maintenance: We will provide reasonable advance notice for planned maintenance windows that may affect gateway availability. Emergency maintenance may be performed without prior notice.

TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, INFYRENCE SHALL NOT BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR PUNITIVE DAMAGES, OR ANY LOSS OF PROFITS, REVENUES, DATA, USE, GOODWILL, OR OTHER INTANGIBLE LOSSES, ARISING OUT OF OR RELATED TO YOUR USE OF OR INABILITY TO USE THE SERVICES.

Our total aggregate liability for any claims arising from or related to the Services shall not exceed the amount you paid us in the twelve (12) months preceding the claim, or $100.00, whichever is greater.

This limitation applies regardless of the legal theory (contract, tort, strict liability, or otherwise) and even if we have been advised of the possibility of such damages.

You agree to indemnify, defend, and hold harmless Infyrence, its officers, directors, employees, contractors, and agents from any claims, damages, losses, liabilities, and expenses (including reasonable attorneys' fees) arising from:

-Your use of the Services, including any inference requests routed through the gateway.-Your violation of these Terms.-Your violation of any rights of a third party, including intellectual property rights.-Any content or data you transmit through the Services.-Your misuse of Virtual Keys or failure to secure your credentials.

You may terminate your account and delete your organization at any time through the dashboard settings or by contacting support. Termination triggers cascading database schema deletions (`onDelete: "cascade"`) of all associated records, including API key hashes, Virtual Keys, telemetry data, and billing meters.

We may suspend or terminate your access to the Services immediately and without prior notice for cause, including but not limited to: violation of these Terms, attempted SSRF attacks, rate limit circumvention, credential abuse, or non-payment.

Upon termination, your right to use the Services ceases immediately. We will make your data available for export for 30 days following termination, after which all remaining data is permanently purged.

These Terms are governed by and construed in accordance with the laws of the State of Delaware, United States, without regard to its conflict of law provisions. Any disputes arising from or relating to these Terms or the Services shall be resolved in the state or federal courts located in Wilmington, Delaware, and you consent to the personal jurisdiction of such courts.

We may update these Terms from time to time to reflect changes in our infrastructure, services, or legal requirements. We will notify you of material changes by posting the updated Terms, updating the "Last updated" date, and sending a notification to the email address associated with your organization.

Your continued use of the Services after any changes constitutes acceptance of the updated Terms. If you do not agree to the modified Terms, you must stop using the Services and delete your organization.