> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tinfoil.sh/llms.txt
> Use this file to discover all available pages before exploring further.

# Safety models

> Safety and content moderation models available on Tinfoil for policy enforcement, guardrails, and Trust & Safety workflows.

> [card]
>
> ![OpenAI](https://mintcdn.com/tinfoil/7kpqELCdP4WIVCil/images/model-icons/openai.png?fit=max\&auto=format\&n=7kpqELCdP4WIVCil\&q=85\&s=1b9cbe9cd3cbfe37d282867e0d448c69)
>
> GPT-OSS Safeguard 120B
>
> [`gpt-oss-safeguard-120b`](#gpt-oss-safeguard-120b)
>
> **Parameters:** 117B (5.1B active)
>
> **Quantization:** MXFP4 (native MXFP4 MoE weights, as released by OpenAI)
>
> **Context:** 131K tokens
>
> **Strengths:** Safety reasoning, bring-your-own-policy flexibility, full access to reasoning chains for debugging, configurable reasoning effort levels
>
> **Structured Outputs:** Structured response formatting support
>
> **Best for:** Content moderation, policy enforcement, LLM guardrails, and Trust & Safety labeling workflows
>
> **Model weights:** [openai/gpt-oss-safeguard-120b](https://huggingface.co/openai/gpt-oss-safeguard-120b)
>
> **Configuration repo:** [tinfoilsh/confidential-gpt-oss-safeguard-120b](https://github.com/tinfoilsh/confidential-gpt-oss-safeguard-120b)
>
> > [!NOTE]
> >
> > **Safety Model:** Classifies text content based on custom safety policies you provide.

> [card]
>
> ![OpenAI](https://mintcdn.com/tinfoil/7kpqELCdP4WIVCil/images/model-icons/openai.png?fit=max\&auto=format\&n=7kpqELCdP4WIVCil\&q=85\&s=1b9cbe9cd3cbfe37d282867e0d448c69)
>
> Privacy Filter
>
> [`pii-filter`](#privacy-filter)
>
> **Parameters:** 1.5B (50M active)
>
> **Strengths:** Token-level PII span detection with typed labels (person, email, phone, address, date, URL, account number, secret), no generative prompting
>
> **Best for:** Detecting and redacting personal information in text, including outgoing web search queries
>
> **Model weights:** [openai/privacy-filter](https://huggingface.co/openai/privacy-filter)
>
> **Announcement:** [Introducing OpenAI Privacy Filter](https://openai.com/index/introducing-openai-privacy-filter/)
>
> **Configuration repo:** [tinfoilsh/confidential-pii-cpu](https://github.com/tinfoilsh/confidential-pii-cpu)
>
> > [!NOTE]
> >
> > Use the authenticated [privacy-filter API](https://docs.tinfoil.sh/guides/privacy-filter.md), or enable `pii_check_options` in [web search](https://docs.tinfoil.sh/guides/web-search.md#pii-protection). The raw endpoint returns the upstream model's spans and placeholder-redacted text; web search separately removes selected spans from outgoing queries. Detection is not a guarantee that all PII will be recognized.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.