> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.fabricgateway.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.fabricgateway.ai/_mcp/server.

# Other LLM providers

> How to connect LLM providers that don't have their own page, including Google Vertex AI.

The console lists each provider that your organization can connect under **LLMs** > **Providers**. This topic covers providers that don't have their own page. For the setup steps, see [Connect an LLM provider](/configure/connections/llm-providers).

## Provider summary

| Provider                                    | API format                  | Credentials                                                                                                               | Notes                                                                                                                                                                      |
| ------------------------------------------- | --------------------------- | ------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Google Vertex AI                            | Set by the provider catalog | A Google Cloud service account key (JSON). On the Enterprise plan, the data plane's own Google Cloud identity also works. | The gateway gets a Google OAuth access token and sends it as `Authorization: Bearer`. The console asks for your **Project ID** and region and puts them in the target URL. |
| Any provider with an OpenAI-compatible API  | OpenAI-compatible API       | Bearer token or API key                                                                                                   | Connect it with a manual connection. See [Connect an OpenAI-compatible or self-hosted model server](/configure/providers/openai-compatible).                               |
| Any provider with an Anthropic Messages API | Anthropic Messages API      | API key                                                                                                                   | Connect it with a manual connection. See [Connect an OpenAI-compatible or self-hosted model server](/configure/providers/openai-compatible).                               |

## What every provider has in common

* The gateway forwards requests in the format the client sent. It doesn't translate between the OpenAI and Anthropic formats.
* The gateway sends the `model` field unchanged. There's no mapping from one provider's model names to another's.
* The gateway adds the provider credential that you set up, and removes the caller's gateway key before it forwards the request.
* A route can combine several connections, even from different providers, as long as they don't declare the same path with different formats. For how to use this, see [Fail over across providers](/configure/routing/failover).

## Next steps

* [Connect OpenAI](/configure/providers/openai)
* [Connect Anthropic](/configure/providers/anthropic)
* [Connect Amazon Bedrock](/configure/providers/amazon-bedrock)