Skip to content
SPECIFICTION
Get listed

Fiddler Guardrails

fiddler.ai

Free plan WebAPISelf-hosted
Visit fiddler.ai
Spec score 8.0 #6 of 23 in AI Guardrail Software

What it is

Fiddler Guardrails is a runtime policy enforcement layer for agentic AI. It evaluates prompts and responses against safety, security, privacy, and accuracy policies, and Fiddler says enforcement takes under 80ms. Safety policies address harmful, inappropriate, and malicious content, including jailbreaks, harassment, hate, illegal content, and violence. Privacy policies detect more than 35 PII entities and healthcare-specific entity types, while faithfulness checks whether responses are grounded in provided context. The service is accessed through an API, with examples for NodeJS, Python, cURL, and HTTP libraries, and includes an out-of-the-box NVIDIA NeMo Guardrails integration. Fiddler says its Centor Models run in customer cloud and VPC environments without data leaving the customer environment. It also states that data is encrypted at rest with AES-256 or higher and communications use HTTPS/TLS 1.2 or higher. The offering has a free plan; Developer is priced on request, billed per trace.

Spec sheet

Purpose
Fiddler Guardrails is a runtime policy enforcement layer for agentic AI that evaluates prompts and responses against safety, security, privacy, and accuracy policies.Read 30 Sep 2026 · fiddler.ai
Compliance
Fiddler states it is SOC 2 Type II audited annually and HIPAA compliant.Read 30 Sep 2026 · fiddler.ai
Support
Enterprise pricing includes white glove support with dedicated communication channels and a named Customer Success Manager.Read 30 Sep 2026 · fiddler.ai
Headquarters
Palo Alto, California, United StatesRead 28 Sep 2026 · fiddler.ai
Founded
2018Read 28 Sep 2026 · fiddler.ai
Accuracy coverage
Faithfulness policies check whether responses are grounded in provided context to prevent misleading output.Read 30 Sep 2026 · fiddler.ai
API and languages
Guardrails is accessed through an API, with code examples for NodeJS, Python, cURL, or any HTTP library.Read 30 Sep 2026 · fiddler.ai
Availability
The documentation describes Guardrails as part of the Fiddler AI Observability and Security platform, accessed in a Fiddler environment using a REST API and API key.Read 30 Sep 2026 · docs.fiddler.ai
Deployment and data
Fiddler says its Centor Models run in customer cloud and VPC environments, with zero data egress from the customer environment.Read 30 Sep 2026 · fiddler.ai
Integration
Fiddler lists an out-of-the-box integration with NVIDIA NeMo Guardrails.Read 30 Sep 2026 · fiddler.ai
Latency
Fiddler states Guardrails enforces policies at runtime in under 80ms.Read 30 Sep 2026 · fiddler.ai
Privacy coverage
Privacy and compliance policies detect 35+ PII entities and healthcare specific entity types for HIPAA compliance, including API keys, payment card numbers, dates of birth, and medical record numbers.Read 30 Sep 2026 · fiddler.ai
Safety coverage
Safety policies block harmful, inappropriate, and malicious content and cover metrics including jailbreaks, harassment, hate, illegal content, and violence.Read 30 Sep 2026 · fiddler.ai
Security controls
Fiddler says customer data is encrypted at rest with AES-256 or higher and communications use HTTPS/TLS 1.2 or higher.Read 30 Sep 2026 · fiddler.ai

Plans and prices

Developer Priced on request per trace Everything in Free · Unified AI observability, including tests and experiments · Custom evaluators and bring your own judge · Role based access control and SSO · SaaS deployment Read 30 Sep 2026 · fiddler.ai
Free Free Real-time guardrails to detect harmful exposure · Protects against hallucinations, toxicity, PII/PHI, prompt injection, and jailbreak attempts · <80ms latency Read 30 Sep 2026 · fiddler.ai
Enterprise Priced on request Everything in Developer · Enterprise-grade guardrails and infrastructure scalability · SaaS, VPC, or on-premise deployment · Dedicated communication channels · Named Customer Success Manager and customized onboarding Read 30 Sep 2026 · fiddler.ai

Every Fiddler Guardrails plan and price →

Strong points and limits

Strong points

  • Runtime policy enforcement is stated as under 80ms.
  • Detects 35+ PII entities and healthcare-specific types.
  • Faithfulness checks responses against provided context.
  • Out-of-the-box NVIDIA NeMo Guardrails integration.
  • Free plan is available.

Limits

  • Access is through an API in a Fiddler environment.
  • Enterprise support details apply to enterprise pricing.

Who it suits

It suits teams building agentic AI applications that need runtime checks for safety, privacy, or grounded responses. API access and the listed deployment options may fit teams integrating guardrails into their own environments.

Verdict

Fiddler Guardrails brings safety, privacy, and accuracy checks to prompt and response handling, with API access and a free plan.

Where Specifiction places it

Alternatives to Fiddler GuardrailsAll alternatives →