---
title: "AI Provider Reference"
slug: "polarity-mcp-ai-providers"
description: "Reference for all six AI provider types supported by Polarity, including configuration requirements and credential setup in the Admin panel."
tags: ["admin panel", "AI provider", "API key", "GPT", "LLM providers", "model settings", "OpenAI", "provider configuration"]
updated: 2026-07-29T18:50:46Z
published: 2026-07-29T18:50:46Z
canonical: "knowledge.threatconnect.com/polarity-mcp-ai-providers"
---

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

# AI Provider Reference

**Polarity supports six AI provider types.**

All provider selection and credential configuration is performed in the “Server Configuration” panel under Model Settings.

| **Provider ID** | **Description** | **Required Config** | **Notes** |
| --- | --- | --- | --- |
| `openai` | OpenAI's GPT model family. | - [ ] API Key | Supports GPT-4o, GPT-4, and GPT-3.5 series. Base URL defaults to https://api.openai.com/v1. |
| `anthropic` | Anthropic's Claude model family. | - [ ] API Key | Supports Claude 3, Claude 3.5, and Claude 4 series. Base URL defaults to https://api.anthropic.com. |
| `gemini` | Google's Gemini model family. | - [ ] API Key | Supports Gemini 1.5 Pro, Gemini 1.0 Pro, and Flash variants. |
| `bedrock` | AWS Bedrock: models served through AWS infrastructure. | - [ ] AWS Access Key ID - [ ] AWS Secret Access Key - [ ] AWS Region - [ ] *optionally* AWS Session Token | Supports Claude, Llama, and Titan models available in Bedrock. No base URL is configurable for this provider. |
| `ollama` | Locally hosted models served via Ollama. | - [ ] Base URL *(e.g., http://localhost:11434)* | No API key required. Suitable for air-gapped environments. Model selection depends on what is available in the local Ollama instance. |
| `openai\_compatible` | Any API that implements the OpenAI chat completions interface. | - [ ] Base URL - [ ] API Key if the endpoint requires it | Compatible with LM Studio, Together AI, Groq, vLLM deployments, and other OpenAI-compatible servers. |

> **Provider credential bypass:**
> 
> The `ollama`, `bedrock`, and `openai\_compatible` providers bypass the standard API key validation path and use their own credential-handling logic.

> [!WARNING]
> Ensure the correct provider-specific fields are filled in before saving.