Legal

Acceptable Use Policy

Rules for safe and lawful use of Switchboard, upstream AI providers, hosted auth, and billing systems.

Production legal documents

Last updated June 8, 2026. Version 2026-06-production-v1.

General rule

Customers may not use Switchboard for unlawful, deceptive, abusive, unsafe, or unauthorized activity, or to help others do so. Customers must maintain appropriate safeguards for their own products and end users.

AI-specific restrictions

Customers may not use Switchboard or upstream AI providers for prohibited AI activity, including child sexual abuse material, illegal harassment, fraud, malware, credential theft, deceptive impersonation, unlawful discrimination, or high-risk decisions without appropriate human review, disclosures, and legal basis.

Payment and billing abuse

Customers may not use Stripe Connect, checkout, billing portals, usage events, refunds, or connected accounts for fraud, transaction laundering, self-funding, evasion of sanctions or payment rules, abusive chargeback behaviour, or activity inconsistent with Stripe terms.

Security misuse

Customers may not probe, scan, attack, overload, bypass rate limits, exfiltrate data, misuse API keys, interfere with isolation between projects, or attempt to access systems, accounts, models, prompts, outputs, billing records, or end-user sessions without authorization.

Spam and communications

Customers may not use Switchboard to send spam, generate deceptive marketing, support unlawful outreach, or violate CASL or other anti-spam rules. Customers are responsible for consent and unsubscribe requirements in their own communications.

Enforcement

Switchboard may investigate, throttle, block, suspend, revoke credentials, disable live billing, preserve evidence, or terminate access when activity appears to violate this policy or creates operational, legal, provider, payment, or security risk.