Audit
SOC 2 Type II via Luxor Technology
Luxor Agent Cloud operates under the SOC 2 Type II program of its parent company. Reports from Luxor Technology are available under NDA upon request.
A SOC 2-aligned program, single-tenant ephemeral VMs, and modern encryption for data at rest and in transit.
Audit
Luxor Agent Cloud operates under the SOC 2 Type II program of its parent company. Reports from Luxor Technology are available under NDA upon request.
Facilities
All Luxor Agent Cloud compute runs in SOC 2-attested data-center facilities with physical access controls, environmental monitoring, and audit logging.
Certification
All Luxor Agent Cloud compute runs in ISO 27001-certified data-center facilities covering information-security management controls.
Isolation
Ensures isolation, strong security and clean execution environments.
Reliability
Uptime SLA, contractually guaranteed and backed by credits.
A SOC 2-aligned program covering audit attestation, a published threat model, and modern encryption for data at rest and in transit.
Luxor Agent Cloud operates under parent company Luxor Technology's SOC 2 Type II program, refreshed annually against the AICPA Trust Services Criteria. Report available under NDA.
Single-tenant, ephemeral VMs isolate every job — fresh kernel, disk, and network.
TLS 1.2+ everywhere, with HSTS enforced and mutually authenticated internal traffic.
AES-256 across storage, databases, and backups, with hardware-backed, rotated keys.
Documented operational practices for detecting, containing, and communicating security events — and a clear path for researchers to report them.
24/7 on-call rotation with a documented IR plan, severity tiers, and post-incident reviews.
24/7 on-callImpacted customers notified within 72 hours of a confirmed incident, with a written post-mortem to follow.
≤ 72 hoursIndependent third-party pen tests at least annually, plus continuous scanning and peer code review.
AnnualReport issues to hello@luxor.ai; we acknowledge valid reports within two business days.
2 business daysLuxor Agent Cloud requests only the GitHub permissions each app needs — nothing more.
Luxor Agent Cloud ships as two separate GitHub Apps (Luxor Agent Cloud for Runners and Luxor Code Reviewer) that are installed independently, and each requests only the scopes it needs. Install one and you never grant the other’s permissions.
Have a question we haven't answered? Reach out to hello@luxor.ai
Luxor Agent Cloud operates under the SOC 2 Type II program of its parent company Luxor Technology. Reports are available under NDA upon request.
All Luxor Agent Cloud compute runs in SOC 2-attested and ISO 27001-certified data centers. Every CI job runs in a fresh, single-tenant VM that is provisioned at job start and destroyed when the job completes — Linux jobs in ephemeral VMs and macOS jobs in isolated VMs on Apple Silicon hardware — so no customer code, dependencies, or secrets persist between runs.
Each job runs in its own fresh, single-tenant VM that is destroyed when the job finishes — on Linux, a hardware-isolated VM with its own kernel and filesystem. There is no shared filesystem, no shared keychain, and no shared user account between jobs. macOS jobs run on Apple Silicon hardware that is multi-tenant at the hardware level, but each job is still isolated in its own VM with no persistent state carried between runs.
Luxor Agent Cloud processes customer data in alignment with the EU General Data Protection Regulation (GDPR). A Data Processing Agreement (DPA) is available on request at hello@luxor.ai. Full GDPR certification is on our compliance roadmap.
Luxor Agent Cloud offers a contractual 99% uptime SLA backed by service credits. SLA terms and credit calculations are defined in the Agent Cloud Master Services Agreement, which we share with prospective enterprise customers during procurement.
Agent Cloud Runners use GitHub Actions' native secret system unchanged. Secrets are injected into the runner VM by GitHub at job start, used by the workflow steps, and destroyed with the VM when the job ends. Luxor Agent Cloud never persists workflow secrets and never sees them outside the lifetime of a single job.
No. Luxor Agent Cloud does not use your code, workflows, or data to train any machine-learning models.
Yes. Report suspected vulnerabilities to hello@luxor.ai with reproduction steps and any supporting artifacts. We acknowledge valid reports within two business days and, once an issue is confirmed, provide an expected remediation timeline — usually within ten business days, depending on severity and complexity. We credit researchers at their request once a fix has shipped. Please don't publicly disclose unfixed issues or test in ways that disrupt other customers.
Yes. SOC 2 Type II reports (via Luxor Technology), Data Processing Agreements, and security questionnaires are all available under NDA. Contact hello@luxor.ai and we will route the request to the Luxor Agent Cloud security and legal team.
Get Luxor Agent Cloud
The cloud for your code and agents.
Set up Sandbox: luxor.ai/agents/sandbox.md