Vertex AI data store - Field content.mime_type must be one of [...] got text/markdown

Google Cloud Build with AI Generative AI & Foundational Models
1

I’m setting up Google Agent Search and followed the documentation to create an unstructured data store backed by a Cloud Storage bucket.

Steps I performed:

  1. Created a Cloud Storage bucket. Uploaded a single Markdown file (.md) to the bucket. The object’s MIME type is text/markdown.
  2. Created an unstructured data store connected to that bucket.

image
image1675×1268 340 KB

The import process starts, but under Data Ingestion Activity the import completes with errors.

image
image2463×818 196 KB

This appears to contradict the official documentation, which states:

Agent Search supports search over documents that are in TXT, JSON, Markdown, PDF, HTML, DOCX, PPTX, XLSX, and XLSM formats.

If Markdown files are supported, why does ingestion fail when the Cloud Storage object has the MIME type text/markdown? Is there an additional configuration required, or is Markdown support not yet available for Cloud Storage ingestion?

I also posted this question on StackOverflow: https://stackoverflow.com/questions/79981288/vertex-ai-data-store-field-content-mime-type-must-be-one-of-got-text-mar.

4

It goes even further. I manually changed the object’s Type in Google Cloud Storage bucket from text/markdown to text/plain.

image
image960×797 51 KB

After waiting for processing to complete, I still received the same warning:

image
image3320×1264 296 KB

Field "content.mime_type" must be one of [application/json, application/pdf, application/vnd.google-apps.document, application/vnd.google-apps.presentation, application/vnd.google-apps.site, application/vnd.google-apps.spreadsheet, application/vnd.ms-excel.sheet.macroenabled.12, application/vnd.openxmlformats-officedocument.presentationml.presentation, application/vnd.openxmlformats-officedocument.spreadsheetml.sheet, application/vnd.openxmlformats-officedocument.wordprocessingml.document, application/xml, go, image/bmp, image/gif, image/jpeg, image/png, image/tiff, text/html, text/plain, text/xml], got "text/markdown".

How can we ingest Markdown files into the Vertex AI data store?