ZNYX AI API

REST API Reference

Every public endpoint exposed by the ZNYX control plane and runtime. Runtime and agentic evaluation, policy authoring, bundles, traces, benchmarks, and evidence - everything you need to ship AI security from code.

Introduction Quick start: evaluate an input

The basics

Authentication

Bearer tokens (runtime, admin, CI) scoped to project + environment. Rotation supported with grace periods.

Runtime evaluate

Low-latency input / output / tool evaluation with streaming support and resolved policy per call.

Policy & bundles

Version, publish, and promote signed policy bundles. Script your CI/CD with a handful of calls.

Errors & rate limits

Typed error envelope, per-plan quotas, and idempotency guidance for mutation endpoints.

Top endpoints by usage

The 80% most developers call in the first week. Start here.

POST/v1/evaluate/input
POST/v1/evaluate/output
GET/v1/bundles/latest
PUT/v1/policies/{tenant}/{app}/{agent}/{env}
POST/v1/orgs/{org}/tokens/runtime
POST/v1/orgs/{org}/projects/{p}/envs/{env}/bundles/publish
POST/v1/orgs/{org}/benchmarks
POST/v1/orgs/{org}/packs/apply

Or browse by category in the sidebar.

All categories

Runtime API

9 endpoints

The hot path — the four endpoints every integration calls before and after the LLM, plus the bundle-fetch endpoint runtimes poll with ETags.

Projects & Environments

18 endpoints

Multi-tenant building blocks. A project groups environments (dev/staging/prod), and each environment owns its own active policy bundle.

Policies

7 endpoints

Read, write, and roll back the detector configuration for a given (tenant / app / agent / env) scope.

Bundles

6 endpoints

Publish, list, activate, and promote signed policy bundles — the deployable artifact runtimes consume.

Traces & Observability

5 endpoints

Query evaluation traces, replay them against newer policy versions, and pull aggregated analytics.

Datasets

9 endpoints

Create custom evaluation datasets, import samples, and browse the built-in ZNYX and industry-standard catalog.

Benchmarks

5 endpoints

Run a dataset against a candidate policy, inspect per-sample results, and diff two runs side-by-side.

Guard Packs

6 endpoints

Browse pre-built vertical policy packs (healthcare, finance, code assistants…) and apply them to a scope.

Output Schemas

7 endpoints

Declare and validate structured-output contracts for AI agents that must emit JSON matching a schema.

Annotations

4 endpoints

Label evaluations as false positive, false negative, correct, or acceptable. Export to a dataset for retraining.

Detector Hub

5 endpoints

Browse and install community or official detectors beyond the built-in set.

Multilingual Packs

8 endpoints

Enable language-specific keyword and pattern overlays (10+ languages) per organization.

Webhooks

6 endpoints

API Keys / Tokens

6 endpoints

Create, list, rotate, and revoke runtime, admin, and CI-scoped API tokens.

Team Management

9 endpoints

Invite, update role, or remove team members. Role-based access control with viewer / editor / admin / owner.

Billing

9 endpoints

Read-only access to plan info, monthly usage, and upgrade flow. Stripe integration for checkout.

Audit Log

2 endpoints

Tamper-evident audit trail for compliance attestations (SOC 2, ISO 27001). Every mutation is hash-chained.

SCIM

14 endpoints

RFC 7644 user-provisioning endpoints for Okta, Azure AD, OneLogin. Enterprise only.

Searching for something specific?

Press ⌘K or Ctrl+K to search every endpoint by name, path, or summary.