Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen3.5, GPT-OSS, Llama, and more!

A User-Focused Photo & File Management System

Adding guardrails to large language models.

The Python error steamroller.

Create Python API documentation in Markdown format.

PλPI

Guardrails AI: PII Filter - Validates that any text does not contain any PII

A test validator repo that includes just the regex validator

FastAPI server for hosting Guardrails AI in your own environment.

A Guardrials Hub validator used to detect if prompt injection is present

Guardrails AI: Profanity free - Validates that a translated text does not contain profanity language.

Adding guardrails to large language models.

📖 Generate markdown API documentation from Google-style Python docstring. The lazy alternative to Sphinx.

Adding guardrails to large language models.

Guardrails AI: Lower case validator - Validates that a value is lower case

Guardrails AI: Valid SQL - Validates that there are no SQL syntactic bugs in the generated code.

Guardrails AI: Redundant sentences - Removes redundant sentences from a string

Guardrails AI: Similar to Document - Validates that a value is similar to the document

A Neural Framework for MT Evaluation

Guardrails AI: QA Relevance LLM eval - Validates that an answer is relevant to the question asked by asking the LLM to self evaluate

Guardrails AI: Excludes SQL predicates - Validates that the SQL query does not contain certain predicates

Guardrails AI: Valid url - Validates that a value is a valid URL

Correct Language Guardrails AI validator - validates whether LLM generated text is in the expected language.

Guardrails AI: Valid JSON - Validates that a value is parseable as valid JSON.

Guardrails AI: Ends with validator - Validates that a list or a string ends with a given value

Guardrails AI: Extracted Summary Sentences Match - Validates that the extracted summary sentences match the original text by performing a cosine similarity in the embedding space

Guardrails AI: Valid length - Validates that the length of value is within the expected range

Guardrails AI: Endpoint is Reachable - Validates that a value is a reachable URL

Guardrails AI: Saliency check - Checks that the summary covers the list of topics present in the document