Guardrails AI
Guardrails AI is an open-source Python framework for defining and applying validators that structure, validate, and constrain generative AI inputs and outputs.
Natural Language Processing in Action, Second Edition
Guardrails AI is an open-source Python package for specifying and validating constraints on large-language-model inputs and outputs.
- How it is used
- Deploys configurable checks as real-time tests around language-model prompts and outputs.
- How deep it goes
- Demonstrates installation and validator configuration, but defers comprehensive attack coverage to classifiers and custom models.
- Examples
- Uses unusual-prompt checks, harmful-output self-checks, and fuzzy spaCy matching for misspelled drug names.
- What you can do after
- Operate/configure validators around model calls and choose exceptions or checks for detected policy violations.
Read how Natural Language Processing in Action, Second Edition explains Guardrails AI →
Machine Learning Platform Engineering
Guardrails AI is a framework for defining and applying validators that check, constrain, and correct generative-model inputs and outputs.
- How it is used
- Places validators before and after LLM processing to screen inputs, prevent leakage, and enforce safety and business policies.
- How deep it goes
- Explains the validator workflow through installation, configuration, concrete checks, failure actions, and Python integration.
- Examples
- Uses competitor-name detection, PII redaction, exception and fix actions, Guardrails Hub installation, and a multilayer safety diagram.
- What you can do after
- Apply it via the Guard API to validate, reject, or sanitize inputs before they reach a production LLM pipeline.
Read how Machine Learning Platform Engineering explains Guardrails AI →
Related concepts
- guardrailuses · A guardrail is a constraint, safeguard, or control that limits actions or behavior to reduce risk and keep a system or process within acceptable bounds.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.