Security

Built for procurement, not around it

Total Shift Left was designed for regulated enterprises that can't ship API specs to OpenAI. This page documents what we ship today, what we're transparent about not yet shipping, and how to get the security questionnaire response your team needs.

Where your data lives

Two deployment shapes — self-hosted on your infrastructure, or multi-tenant SaaS managed by us. Both shapes treat your API specifications, credentials, and test traffic as sensitive by default.

Data classSelf-hosted deploymentMulti-tenant SaaS
API specifications (OpenAPI / Swagger / WSDL)Stays on your infrastructureStored in tenant-isolated MongoDB
AI prompts and generated testsSent only to the LLM endpoint you configure (Ollama / vLLM / LM Studio)Sent to the LLM provider you select (BYO API key)
API credentials, tokens, secretsAES-256-CBC encrypted at rest in your databaseAES-256-CBC encrypted at rest, tenant-isolated
Test execution traffic (HTTP requests to your APIs)Originates from your runner; never leaves your networkOriginates from cloud runner; egress to your published API endpoints only
Audit logsStored in your database; configurable retention on roadmapStored per-tenant; export available via API

For deployment topology and runtime requirements, see the deployment page.

Identity, access, and audit

Role-based access control

Five built-in roles cover the typical enterprise testing org: Administrator, Contributor, Reviewer, Reader, and Environment Manager. Roles map to project assignments and govern who can author tests, approve API spec changes, run executions, and view results.

Authentication

Local authentication with JWT tokens, AES-256 encrypted credential storage, and email-based user invitations. Token expiry, role-based scoping, rate limiting, and CORS controls protect the public REST API.

Single sign-on

Azure AD / Entra ID, generic OIDC (Okta, Google Workspace), and SAML 2.0 are on the near-term roadmap, with auto-provisioning and group-to-role mapping. See the SSO row on the pricing comparison for current status. Enterprise customers can integrate via professional services in the interim.

Audit & observability

Audit logs capture every meaningful action with role context, filterable viewer, statistics, and export. Real-time logs and a system health dashboard surface platform metrics. Configurable retention is on the roadmap; activity feeds are scheduled to ship alongside.

AI governance: your specs, your LLM, your call

The single biggest reason regulated buyers move off cloud-only API testing tools: the AI features require shipping API specifications to a third-party LLM. We treat that as a configuration choice, never a default.

BYO LLM key

Cloud LLM access is always bring-your-own-key. We never broker credentials to OpenAI, Anthropic, Azure OpenAI, Gemini, or any of the 13+ supported providers. Your key, your usage, your invoice.

Self-hosted inference (Enterprise)

Run AI test and mock generation against an Ollama, vLLM, or LM Studio endpoint inside your perimeter — or any OpenAI-compatible URL you control. API specifications, prompts, and generated tests are sent only to the endpoint you configure.

Spec change approval

When an upstream API spec changes, the platform flags affected tests and routes the change through an explicit review and approval flow before regenerating tests. Schema drift is surfaced, not silently absorbed.

Compliance posture

Honest about what is shipped, what is on the roadmap, and what is configurable today. We will update this page as items move state.

  • SOC 2

    Roadmap

    Engagement with an independent auditor is required before publishing a Type 1 target date. We will update this page when that engagement begins.

  • ISO 27001

    Roadmap

    Aligned with internal security governance; formal certification follows SOC 2.

  • GDPR / data residency

    Self-hosted by default

    Self-hosted deployment keeps API specs, credentials, and test data inside the residency boundary you choose. Multi-tenant SaaS region selection is on the roadmap.

Security questionnaire response

Procurement and security teams typically ask for the same 30-50 control questions. We share a security questionnaire response, deployment topology diagram, and reference architecture upfront on the architect call — so your security team can review in parallel with the technical evaluation.

Email security@totalshiftleft.com with your questionnaire format, or book a demo and we'll share the standard response packet on the call.

Talk to our architect, not a sales rep

30-minute deployment and security review. We share questionnaire responses, topology diagrams, and reference architecture on the call.