Vector indexing takes forever and vector_stores.files.list returns incorrect results

API
,
1

Hello,

I have two issues with vector stores (these issues are potentially related).

First, on the Dashboard, my vector store has 0 file:

Screenshot 2025-11-06 at 18.26.15
Screenshot 2025-11-06 at 18.26.151366Ă—1200 63.2 KB

Yet when querying using the API, it has one file (`in_progress`):

import asyncio
import logging

from openai import AsyncOpenAI

client = AsyncOpenAI()

async def test_vs_file_list():
    vector_store_id = "vs_690392abfb6881919b48c074872cb5d6"
    paginator = client.vector_stores.files.list(vector_store_id=vector_store_id)
    async for item in paginator:
        logging.info(f"Vector store file: {item}")

async def main():
    await test_vs_file_list()
    

asyncio.run(main())

The above code gives:

2025-11-06 18:22:43,089 - INFO - HTTP Request: GET https://api.openai.com/v1/vector_stores/vs_690392abfb6881919b48c074872cb5d6/files "HTTP/1.1 200 OK"
2025-11-06 18:22:43,102 - INFO - Vector store file: VectorStoreFile(id='file-JLj86gi4qwW29BDyWUc2U9', created_at=1761841905, last_error=None, object='vector_store.file', status='in_progress', usage_bytes=8667749, vector_store_id='vs_690392abfb6881919b48c074872cb5d6', attributes={}, chunking_strategy=StaticFileChunkingStrategyObject(static=StaticFileChunkingStrategy(chunk_overlap_tokens=400, max_chunk_size_tokens=800), type='static'))

Second issue: As you can see from the above logs, the file was uploaded on October 30th (i.e., a week ago), yet it’s still “in_progress”. Is this a bug?

Thank you in advance for your help!

1 Like
2

I forgot to mention that the file is quite big, 39.8 MB.

3

Docs (only under “file search and retrieval” → “retrieval”)

The maximum file size is 512 MB. Each file should contain no more than 5,000,000 tokens per file (computed automatically when you attach a file).

If English text was extracted, you’ll get a compression ratio about 4:1, so you may exceed the token count at 15MB - 20MB. Non-latin languages are worse.

The ratio and tolerance for file size may be better if a PDF or other binary with lots of images that have no search text.

The platform site code may only display completed status files. It uses a list method similar to the API, if you were to monitor network requests with browser developer tools and want to see if OpenAI’s code is getting a different files listing.

“Last active” was two minutes after creation. You can either delete the file attachment, or delete the vector store ID and try again.

4

Is this the endpoint that you’re referring to?

https://platform.openai.com/docs/api-reference/vector-stores-files/listFiles

Overall, it seems that that endpoint is very buggy. I posted a related issue:
Vector Store List Files Bug - Deleted files show as In Progress - #3 by olivier.cuny

What do you get when doing query with this endpoint:

https://platform.openai.com/docs/api-reference/vector-stores-files/getFile

Are you still getting a status of “in_progress” or it is simply not finding your file?

1 Like