Notifications
Clear all
Topic starter
I have cleaned text content (from scraping) which I want to get text embeddings for. The input text could come from BigQuery or Google Sheets, and I want to use a Google Vertex embedding function in callin.io.
According to the callin.io integrations page, the Google Vertex AI Embeddings node is available as a regular node:
However, in my self-hosted version of callin.io, this node is not available in the node selector:
Interestingly, when I try the callin.io cloud version, the Vertex AI Embeddings integration does appear, but it’s available as a “tool”, not a typical workflow node - it doesn’t connect to other nodes.
Is Google text embedding (and other vertex AI features) not available as in-built nodes? Or Am I missing something here?
P.S.
I’m aware that I can use http request to call the Google Cloud Console APIs, but I want to use native/in-built nodes if available.
Integrations mentioned:
Information on your callin.io setup
- callin.io version: 1.88.0
- Running callin.io via (Docker, npm, callin.io cloud, desktop app): Self hosted (npm)
- Operating system: Windows 11 64-bit
Please provide the rewritten markdown content *it should be in the markdown format.
Posted : 17/04/2025 2:35 am
Hello!
The embeddings tool nodes are designed to be linked with AI agent nodes or other nodes that support automatic invocation of the embeddings tool, such as vector store nodes.
What was your intended use for the embeddings data within your workflow? For example, if you are aiming to build an integration from BigQuery to a vector database, you could utilize Google Vertex embeddings by constructing something similar to this:
That integrations page appears somewhat misleading, as we do not currently feature a standalone embeddings node as depicted in the screenshot. I will escalate this matter.
Posted : 17/04/2025 8:10 am
Topic starter
Hi there!
My goal is to compute similarity scores between various documents I've collected, utilizing text embeddings. Based on your feedback, it seems I'll need to use custom HTTP requests to interact with text embedding APIs directly.
Yes, I noticed a few similar cases in the integrations section where the functionality wasn't immediately obvious. These might benefit from a review.
Thanks for the help!
Posted : 17/04/2025 5:01 pm
This discussion was automatically closed 90 days following the last response. New replies are no longer permitted.
Posted : 16/07/2025 5:01 pm