The native Google Vertex AI resource is now functional, but its model selection only exposes Gemini models. The request is to add support for Vertex AI partner models β particularly Anthropic Claude models available through Vertex AI Model Garden β since accessing them via a custom REST resource forfeits native features like Retool Vectors.
AI Response
This behavior is expected: the native Google Vertex AI resource is currently scoped to the Gemini family of models only. Retool's documentation explicitly lists the supported models (Gemini 2.5 Flash/Pro and the Gemini 3.x models) and describes the integration as leveraging "Gemini's capabilities," with no support for Vertex AI Model Garden partner models such as Anthropic Claude. There is no native option today to select Claude-via-Vertex within this resource, so a custom REST resource remains the only route β with the noted trade-off of losing native AI features like Retool Vectors. Adding partner-model support would be a feature request rather than something configurable today.
Sources
Connect to Google Vertex AI | Retool Docs
The official connect guide confirms the native Vertex AI resource supports only the Gemini model family, with a Supported Models table that lists no partner models like Claude. Google Vertex AI
The product update announcing the native Vertex AI integration frames it around Gemini models in AI resource queries, corroborating the Gemini-only scope.
The Community Team is testing out a new automation. Let us know if it's helpful (or not) by leaving a , , or . Or by marking this post as the "Solution"! Let us know if you have any feedback here.
Hi @Jenna_Hilden! Looks like there is an internal feature request for this (that @Darren has filed) to include partner models.
The current workaround is to use a RestAPI resource to integrate with unsupported models, but understandable that this does not suit your needs at this time. Standing up an OpenAI-compatible proxy in front of Vertex and use Custom Provider would also work as well.
There hasn't been much of an update with the feature request, but I have just bumped the request and will be sure to circle back here as soon as I get one!