"Upload has already been terminated" error with markdown files after Gemini File Search API update

Gemini API
,
1

I’m experiencing an issue with the File Search API after the recent update (4 days ago). https://ai.google.dev/api/file-search/file-search-stores, Uploading markdown (.md) files fails with:

400 Bad Request: {‘message’: ‘Upload has already been terminated.’, ‘status’: ‘Bad Request’}

Additionally, I see 503 ServiceUnavailable errors in AI Studio.

What works:

  • PDF file uploads work fine
  • The uploadToFileSearchStore method executes

What fails:

  • Markdown file uploads consistently fail
  • Operation tracking throws “already terminated” error

Root cause I discovered:
The API documentation shows a new endpoint structure:

  • Old: fileSearchStores/{store}/operations/{id}
  • New: fileSearchStores/{store}/upload/operations/{id} (for upload operations)

However, the Python SDK (google-genai) doesn’t have client.file_search_stores.upload.operations yet - it only has client.operations.get() which uses the old endpoint.

Questions:

  1. When will the Python SDK be updated to support the new /upload/operations/ endpoint?
  2. Is there a workaround besides making direct REST API calls?
  3. Why do PDFs work but markdown files don’t with the same upload method?

Environment:

  • Python SDK: google-genai (latest)
  • Files: Markdown (.md) with explicit mime_type: ‘text/markdown’

Any guidance would be appreciated!

1 Like
Gemini File Search API - "Upload has already been terminated" error when uploading large CSV files (>5MB)
3

Hi @0xneobyte
Welcome to the AI Forum!!!

Thank you for bringing this to our attention.
Could you please share the full payload details along with a sample of the code that you are using? We would like to reproduce the issue.

5

Environment:

  • Python SDK: google-genai (latest version from pip)
  • Python: 3.11+
  • OS: macOS arm

Full Code Sample:

from google import genai
import time

client = genai.Client(api_key="YOUR_API_KEY")

# This works fine for PDFs
pdf_operation = client.file_search_stores.upload_to_file_search_store(
    file="/path/to/document.pdf",
    file_search_store_name="fileSearchStores/YOUR_STORE_ID",
    config={
        'display_name': 'Test PDF',
        'custom_metadata': [
            {"key": "subject", "string_value": "mathematics"}
        ],
        'mime_type': 'application/pdf'
    }
)

works fine for PDF uploads

while not pdf_operation.done:
    time.sleep(2)
    pdf_operation = client.operations.get(pdf_operation)

print("PDF uploaded successfully")

This fails for markdown files

md_operation = client.file_search_stores.upload_to_file_search_store(
    file="/path/to/document.md",
    file_search_store_name="fileSearchStores/YOUR_STORE_ID",
    config={
        'display_name': 'Test Markdown',
        'custom_metadata': [
            {"key": "subject", "string_value": "mathematics"}
        ],
        'mime_type': 'text/markdown'
    }
)

ERROR

# Wait for markdown operation - FAILS HERE
while not md_operation.done:
    time.sleep(2)
    md_operation = client.operations.get(md_operation)  # Throws error

print("Markdown uploaded successfully")

Error Output:
400 Bad Request: {'message': 'Upload has already been terminated.', 'status': 'Bad Request'}

Key Observations:

  1. PDFs work perfectly with the exact same code pattern
  2. Only markdown files consistently fail
  3. The new API documentation (updated 4 days ago) shows upload operations should use: fileSearchStores/{store}/upload/operations/{id}
  4. However, the SDK’s client.operations.get() uses the old endpoint: fileSearchStores/{store}/operations/{id}
  5. I verified the SDK doesn’t have client.file_search_stores.upload.operations namespace yet
  6. this code worked well before (like 2-3 weeks back
6

WAIT, NOW IT WORKS!!!

i think there was something wrong in the backend and now its fixed! thanks.

7 
I've been facing the same issue since its release. It happens with some PDF files, not all.

File: some PDFs, not all

operation = self.client.file_search_stores.upload_to_file_search_store(
    file=file_obj,
    file_search_store_name=f"fileSearchStores/{settings.gemini_file_store_id}",
    config=upload_config
)

The process stalls for a while then returns:

400 Bad Request: {'message': 'Upload has already been terminated.', 'status': 'Bad Request'}
9

are you fix this prob?
facing !