Legal
Privacy Policy
Overview
Swift Inference Systems Inc. ("SwiftInference," "we," "us," or "our") operates an edge AI inference platform that deploys GPU compute at cellular tower and carrier POP sites. This Privacy Policy explains how we collect, use, disclose, and protect information about you when you use our website, platform, APIs, and services (collectively, the "Services").
By using our Services, you agree to the practices described in this policy. If you do not agree, please discontinue use of the Services.
Information we collect
Account & identity information
When you register for an account or request a pilot, we collect:
- Name, work email address, and company name
- Role and contact preferences
- Billing and payment information (processed by our payment processor; we do not store full card numbers)
- API keys and authentication credentials you generate within the dashboard
Usage & telemetry data
We automatically collect operational data when you interact with our platform, including:
- API request metadata: timestamps, endpoint paths, response codes, token counts, and latency measurements (p50, p90, p99, TTFT)
- Node and slot utilization metrics (GPU memory, throughput, error rates)
- Dashboard interactions and feature usage
- IP address, browser type, device information, and referring URLs for web access
| Data category | Examples | Purpose |
|---|---|---|
| Identity | Name, email, company | Account management, communications |
| Billing | Token counts, slot usage | Accurate invoicing and attribution |
| Telemetry | Latency, error rates, GPU util | Platform reliability and SLA reporting |
| Logs | Request IDs, node IDs, timestamps | Debugging, security, compliance |
| Cookies | Session tokens, preferences | Authentication, UX consistency |
Information you provide voluntarily
We collect information you actively provide, such as pilot request forms, support tickets, partnership inquiries, and feedback submitted through the dashboard or email.
How we use your information
We use collected information to:
- Operate the platform — authenticate users, route inference requests, enforce slot assignments, and manage node health
- Bill accurately — meter token consumption per customer and model, generate invoices, and attribute usage
- Monitor performance — track SLA compliance, latency distributions, error rates, and GPU utilization across edge nodes
- Improve the Services — analyze aggregated, de-identified usage patterns to prioritize engineering work and optimize routing logic
- Communicate with you — send transactional emails (account changes, invoices, incident alerts) and product updates you opt into
- Ensure security — detect abuse, investigate anomalies, and protect the integrity of our infrastructure and customers' workloads
- Comply with law — respond to valid legal processes and meet regulatory obligations
We do not use your information for advertising, profiling unrelated to the Services, or any purpose incompatible with those listed above without your explicit consent.
Inference & model data
Inference payload handling
Inference requests flow directly to edge nodes over private on-net paths. The control plane coordinates routing but does not proxy or log the content of inference payloads in steady-state operation. Specifically:
- Prompt and completion text is not stored beyond the active request lifetime unless you explicitly enable logging features in your account settings
- Audio transcription inputs are processed in memory and discarded after the response is returned
- We log request metadata (token counts, latency, node ID, customer ID) for billing and observability — not payload content
Customer-uploaded models
Model artifacts you upload are stored in isolated, access-controlled object storage. They are used exclusively to serve your inference workloads and are never accessed, copied, or disclosed to other customers or third parties. You retain full ownership of your model artifacts and may delete them at any time.
WireGuard & management traffic
Management traffic between edge nodes and the control plane (heartbeats, metrics, deployment commands) travels over encrypted WireGuard tunnels. This traffic is operational in nature and does not contain inference payload content.
Data sharing
We do not sell your personal information. We share data only in the following limited circumstances:
Service providers
We engage trusted third-party vendors who process data on our behalf under strict confidentiality obligations. These include cloud infrastructure (AWS us-west-2 for the control plane), payment processing, and email delivery. Vendors receive only the data necessary for their function.
Infrastructure partners
Tower and carrier partners (such as American Tower and AT&T) provide physical rack space, power, and network connectivity. These partners receive no access to your account data, inference payloads, or model artifacts. Network-level metadata such as IP addresses may be visible to carriers as a consequence of routing traffic through their infrastructure.
Legal requirements
We may disclose information where required by law, subpoena, or valid legal process, or to protect the rights, property, or safety of SwiftInference, our customers, or the public. We will notify you of such requests to the extent permitted by law.
Business transfers
If SwiftInference is acquired, merges with another entity, or transfers assets, your information may be transferred as part of that transaction. We will provide notice before your data becomes subject to a materially different privacy policy.
Data retention
We retain different categories of data for different periods:
| Data type | Retention period | Rationale |
|---|---|---|
| Account information | Duration of account + 90 days | Account management and legal obligations |
| Billing records | 7 years | Tax and financial compliance (append-only) |
| Request metadata logs | 30 days (Loki retention) | Debugging and SLA verification |
| Metrics (Prometheus) | 13 months | Year-over-year performance analysis |
| Inference payloads | Not retained (in-memory only) | Privacy by design |
| Model artifacts | Until customer deletion | Customer ownership and control |
| Support correspondence | 3 years | Service continuity and dispute resolution |
When you close your account, we delete or anonymize your personal information within 90 days, except where retention is required by law or necessary to resolve outstanding disputes.
Security
We implement layered technical and organizational safeguards to protect your data:
- Encryption in transit — all external communications use TLS 1.2+; inter-node management traffic uses WireGuard (ChaCha20-Poly1305)
- Encryption at rest — model artifacts and databases are encrypted at rest using AES-256
- Node attestation — edge nodes are registered with signed certificates; private keys never leave the node
- Secure boot & signed updates — node software is verified before execution
- API key scoping — keys are scoped per customer and can be revoked instantly
- Access controls — internal access to customer data is restricted to personnel with a legitimate need and logged
- Inbound-dark nodes — edge nodes accept no unsolicited inbound connections; all management is outbound-initiated
No system is perfectly secure. If you discover a security vulnerability, please contact us at security@swiftinference.ai before public disclosure.
Your rights
Depending on your location, you may have the following rights regarding your personal information:
- Access — request a copy of the personal data we hold about you
- Correction — update or correct inaccurate information via the dashboard or by contacting us
- Deletion — request deletion of your account and associated personal data (subject to legal retention requirements)
- Portability — receive your data in a machine-readable format
- Restriction — ask us to restrict processing of your data in certain circumstances
- Objection — object to processing based on legitimate interests
- Opt-out of marketing — unsubscribe from non-transactional communications at any time via the unsubscribe link or by contacting us
To exercise any of these rights, email us at privacy@swiftinference.ai. We will respond within 30 days. We may need to verify your identity before fulfilling your request.
California residents may have additional rights under the California Consumer Privacy Act (CCPA). We do not sell personal information as defined by the CCPA. For CCPA requests, use the same contact above.
Cookies & tracking
We use a minimal set of cookies and similar technologies on our website and dashboard:
- Strictly necessary cookies — session authentication tokens required to maintain your logged-in state. These cannot be disabled without breaking the Services.
- Preference cookies — store dashboard UI preferences (theme, table sort order, etc.) so your settings persist across sessions.
- Analytics — we may use privacy-respecting, first-party analytics to understand aggregate page usage. We do not use third-party behavioral advertising trackers.
You can control cookies through your browser settings. Disabling strictly necessary cookies will prevent you from using authenticated features of the Services.
Children's privacy
The Services are designed for businesses and professional developers. We do not knowingly collect personal information from individuals under the age of 16. If we learn we have collected data from a child under 16, we will delete it promptly. If you believe a minor has provided us information, contact us at privacy@swiftinference.ai.
Changes to this policy
We may update this Privacy Policy from time to time. When we make material changes, we will:
- Update the "Last updated" date at the top of this page
- Send an email notification to account holders for significant changes
- Display a notice in the dashboard for 30 days following the change
Continued use of the Services after the effective date of an updated policy constitutes acceptance of the changes. We encourage you to review this page periodically.
Contact us
If you have questions, concerns, or requests related to this Privacy Policy or your personal data, please reach out: