LLM Guard

license

MIT(OSI)

source

public(full Python implementation)

self-host

yes

service

none(the OSS toolkit

core features withheld

no

LLM Guard is released under the MIT license with its complete Python source, and it installs from PyPI or runs as a self-hosted API. None of its documented scanners depends on an enterprise build, a license key or a hosted service.

• https://github.com/protectai/llm-guard recorded 2026-08-13

  Repo metadata records license spdxId - MIT. README calls it 'an open source solution', documents installing and customizing it yourself and deploying it as an API, and enumerates the prompt and output scanners in the published package; there is no enterprise directory, commercial edition or license key on the page.

• https://protectai.github.io/llm-guard/ recorded 2026-09-29

  The LLM Guard documentation site lists tutorials for "Secure RAG with Langchain" and "Secure Agents with Langchain", input and output scanner pages, and API pages including "Deployment" and "API Reference", and repeats that "LLM Guard is an open source solution."

• https://pypi.org/pypi/llm-guard/json recorded 2026-09-29

  The PyPI JSON for the llm-guard package (version 0.3.16) carries the full license text, which opens "The MIT License (MIT)" with "Copyright (c) Protect AI. All rights reserved.", and lists its repository as "https://github.com/protectai/llm-guard". The embedded README says "LLM Guard is an open source solution", installs with "pip install llm-guard", notes "Base functionality requires a limited number of libraries" and links a guide to "Deploy LLM Guard as" an API. Searched the body for enterprise, license key, commercial and on-prem: none.

Adoption is read from PyPI downloads of the llm-guard package, confirmed as the product's own release by its project URLs pointing back to protectai/llm-guard. PyPI is the primary distribution channel for a Python library, so its downloads are used here rather than GitHub stars.

• https://pypistats.org/api/packages/llm-guard/recent recorded 2026-08-13

  last_month 212,987

LLM Guard offers a broad set of input and output scanners that are easy to drop into an existing pipeline, but it has no policy language or dialog control of its own. That places it a step below NeMo Guardrails, which adds both on top of a similar scanner set.

• https://github.com/protectai/llm-guard recorded 2026-08-13

  README lists the prompt scanners (Anonymize, BanCode, BanCompetitors, BanSubstrings, BanTopics, Code, Gibberish, InvisibleText, Language, PromptInjection, Regex, Secrets, Sentiment, TokenLimit, Toxicity) and the output scanners (BanCode, BanCompetitors, BanSubstrings, BanTopics, Bias, Code, Deanonymize, JSON, ...), and frames the toolkit as sanitization, detection of harmful language, prevention of data leakage and resistance to prompt injection.

• https://pypi.org/pypi/llm-guard/json recorded 2026-09-29

  The PyPI README for llm-guard lists prompt scanners including "PromptInjection", "Anonymize", "Secrets" and "Toxicity", and output scanners including "Deanonymize", "MaliciousURLs", "FactualConsistency" and "Sensitive", and describes the package as "offering sanitization, detection of harmful language, prevention of data leakage, and resistance against prompt" injection attacks.