OpenGuardrails

What OpenGuardrails is

OpenGuardrails AIRS is a commercial runtime service for testing, inspecting, and protecting AI agents. It evaluates tool calls, resources, and accounts before a model sees a request, then checks responses before an agent acts. Runtime decisions can allow, block, or redact, with findings and redaction spans. The console includes a live monitor, agent and session explorer, findings triage, policies, a playground, and API keys. Integrations use the API directly rather than an SDK; the documented recipe uses two POSTs per model call. Starter masks secrets on the customer host before requests leave, then restores them only into tool arguments. Private deployments can run on customer Kubernetes or VMs with customer datastores and a model gateway in the customer environment. The OGR protocol is an open Apache-2.0 specification with published JSON Schemas, while the runtime product is commercial. Starter costs $10 a month, billed monthly with cancellation anytime; there is no free plan. The default rate limit is 600 requests per minute per API key. If the runtime is unreachable, local fail policies default to open unless categories are explicitly set to closed.

Who it suits

OpenGuardrails suits teams that need runtime checks for AI agents and can integrate through an API. Its private deployment option is relevant to customers wanting infrastructure, datastores, and a model gateway in their own environment.

Limits

Starter costs $10 a month and includes secret masking, a secret ledger, and detection for secrets, PII, dangerous commands and credential files. Pro costs $20 a month and adds model judges, one-agent sizing and four policy templates; the default API-key rate limit is 600 requests per minute.

Getting started

1. Choose a Starter, Pro, Max or Enterprise plan
2. Use the API directly because integrations do not use an SDK
3. Configure the runtime and local fail policies for unreachable-service behavior
4. Connect a listed integration or your own API workflow
5. For private deployment, run it on customer Kubernetes or VMs with customer datastores

Verdict

OpenGuardrails offers runtime agent checks, a monitoring console, and private deployment options, with a commercial runtime and an open protocol specification. Teams should consider its API integration approach, rate limit, and fail-policy behavior when evaluating it.