sure, linked, name - sharprecap-article-embeddings-prod

DDmitri sure, linked, name - sharprecap-article-embeddings-prod

D

DmitriOP•11/19/25, 7:30 PM

would also add - graphs from smaller index (±3400 vectors) - its responsive, works...
and from bigger one - 1 mln+ - should be ±30k queries on display, shows literally nothing...
worked like a charm before yesterday...

DDmitri sure, linked, name - sharprecap-article-embeddings-prod

P

Peps•11/19/25, 9:32 PM

I've asked someone of the team to take a look at your index

E

Egor•11/24/25, 8:56 AM

For AutoRAG's 'Max files per AI Search' limit of 100k files: does this mean AutoRAG stops indexing after 100k files are added to the R2 bucket, or does it mean only 100k files are searchable at any given time? If it's the latter, could I work around this in a multitenant app by scoping searches to specific user folders via metadata filters?

https://developers.cloudflare.com/ai-search/platform/limits-pricing/#limits

DDmitri Could anybody advice - how to get help or support with vectorize? I don't see an...

P

peterp•11/24/25, 9:13 PM

We have the same issue... Did you get this resolved?

Ppeterp We have the same issue... Did you get this resolved?

P

peterp•11/25/25, 10:16 AM

We created a new index and that resolved things. We would love to know what happened.

V

Vlad Lykhonis•11/28/25, 3:39 PM

Hm, I am calling VectorizeIndex.describe() which should return:

interface VectorizeIndexDetails {
  /** The unique ID of the index */
  readonly id: string;
  /** The name of the index. */
  name: string;
  /** (optional) A human readable description for the index. */
  description?: string;
  /** The index configuration, including the dimension size and distance metric. */
  config: VectorizeIndexConfig;
  /** The number of records containing vectors within the index. */
  vectorsCount: number;
}

interface VectorizeIndexDetails {
  /** The unique ID of the index */
  readonly id: string;
  /** The name of the index. */
  name: string;
  /** (optional) A human readable description for the index. */
  description?: string;
  /** The index configuration, including the dimension size and distance metric. */
  config: VectorizeIndexConfig;
  /** The number of records containing vectors within the index. */
  vectorsCount: number;
}

interface VectorizeIndexDetails {
  /** The unique ID of the index */
  readonly id: string;
  /** The name of the index. */
  name: string;
  /** (optional) A human readable description for the index. */
  description?: string;
  /** The index configuration, including the dimension size and distance metric. */
  config: VectorizeIndexConfig;
  /** The number of records containing vectors within the index. */
  vectorsCount: number;
}

interface VectorizeIndexDetails {
  /** The unique ID of the index */
  readonly id: string;
  /** The name of the index. */
  name: string;
  /** (optional) A human readable description for the index. */
  description?: string;
  /** The index configuration, including the dimension size and distance metric. */
  config: VectorizeIndexConfig;
  /** The number of records containing vectors within the index. */
  vectorsCount: number;
}

but instead I get only seems like this part:

{
  "dimensions": 768,
  "vectorCount": 0
}

{
  "dimensions": 768,
  "vectorCount": 0
}

V

Vivek•12/2/25, 11:51 AM

Is anyone else seeing metadata indexes not being created in Cloudflare Vectorize?
Sometimes the index creation never finishes, and using

wait=true

wait=true

to poll the mutation ID either throws 504 or takes a very long time.
Any workaround to make index creation reliable

VVivek Is anyone else seeing metadata indexes not being created in Cloudflare Vectorize...

P

Priyanshu Gupta•12/2/25, 12:59 PM

For me its taking indefinite time some times, making it highly unreliable for metadata index creation via cloudflare vectorize

VVivek Is anyone else seeing metadata indexes not being created in Cloudflare Vectorize...

E

Egor•12/7/25, 2:57 PM

Yep. I switched off the whole AI search and vectorize. It is too unreliable at this point

sure, linked, name - sharprecap-article-embeddings-prod

Similar Threads

Similar Threads

Similar Threads