Related Issues

• partially addresses SupabaseBucketDownloader #3084

Proposed Changes

Adds SupabaseBucketDownloader, a new Haystack component that downloads files from a Supabase Storage bucket and returns them as ByteStream objects for use in indexing pipelines.

Component: haystack_integrations.components.downloaders.supabase.SupabaseBucketDownloader

Key features:

• Downloads files from a Supabase Storage bucket by file path
• Returns ByteStream objects with meta["file_path"] and meta["bucket_name"] set
• Optional file_extensions filter (case-insensitive)
• Failed downloads are logged and skipped — does not raise, returns remaining streams
• MIME type auto-detected from file extension
• Full serialization support (to_dict / from_dict)
• Authenticates via SUPABASE_SERVICE_KEY env var (service role key for private buckets)

Usage:

from haystack_integrations.components.downloaders.supabase import SupabaseBucketDownloader
from haystack.utils import Secret

downloader = SupabaseBucketDownloader(
    supabase_url="https://<project-ref>.supabase.co",
    supabase_key=Secret.from_env_var("SUPABASE_SERVICE_KEY"),
    bucket_name="my-documents",
)
result = downloader.run(sources=["reports/report.pdf", "data/notes.txt"])
streams = result["streams"]

Dependency added: supabase>=2.0.0

How did you test it?

9 unit tests covering:

• Default and custom init parameters
• File extension normalization (case-insensitive)
• Serialization round-trip (to_dict / from_dict)
• run() returns correct ByteStream objects with metadata
• File extension filtering skips non-matching files
• Empty sources returns empty list
• Failed downloads are skipped (logs warning, continues)
• MIME type detection from file extension

Integration test skipped unless SUPABASE_URL and SUPABASE_SERVICE_KEY are set.

Notes for the reviewer

This is the first of the remaining sub-tasks from #2862 (after SupabasePgvectorDocumentStore, SupabasePgvectorEmbeddingRetriever, and SupabasePgvectorKeywordRetriever were merged in #3164).

The supabase>=2.0.0 dependency is added to pyproject.toml. The existing integration only depended on pgvector-haystack.

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added unit tests and updated the docstrings
• I've used one of the conventional commit types for my PR title: feat: