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

# AI assistant

> Ask about your findings, what's overdue and what to fix first, or get help reviewing code and remediating an issue. Bring your own key, or use a model you run yourself.

## Overview

The AI assistant opens in a side panel from anywhere in Tierity. It can read your organization's findings and connected tools, so you can ask questions about your own security program instead of writing queries.

Click **AI Assistant** in the sidebar to open it.

## What to ask

> **Triage**
>
> "What's overdue?", "Which critical findings are newest?", "What should I fix first this week?"

> **Remediation**
>
> "How do I fix this finding?", "Draft a ticket for the Okta admins without MFA."

> **Code review**
>
> Paste code and ask for security issues, insecure patterns or misconfigurations.

> **Research**
>
> CVEs, vulnerability classes, library risks and how an attack technique works.

## Organization instructions

Admins can give the assistant standing context about your organization under **Config → General**, for example "We're a fintech company; always flag PCI-DSS relevant findings. Our main stack is Python, FastAPI and PostgreSQL."

## Providers

The assistant is bring-your-own-key on every plan. Add a provider under **Config → AI Providers** (Manage role or higher). On self-hosted instances, a server admin can also set a server-wide default; see [AI providers](https://docs.tierity.com/admin/ai-providers.md).

| Provider | Notes |
| - | - |
| **Anthropic** | Claude models, with your Anthropic API key |
| **OpenAI** | GPT models, with your OpenAI API key |
| **Google Gemini** | Gemini models, with a Google AI Studio key |
| **Ollama** | Models you run yourself; no key needed |
| **OpenAI-compatible** | Any server that implements the OpenAI chat completions API, such as vLLM or LM Studio |

## Data privacy

> [!WARNING]
>
> Questions, and the findings the assistant reads to answer them, are sent to the AI provider whose key you configured, under that provider's data terms. Agree internally on what can be shared, and don't paste credentials into the chat.

Because the key is yours, requests go straight to your own provider account.


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