Roman

@evgenydmitriev , I've added unique text processing and updated tests. Only unique texts are sent to workers, but results are assigned back to original inputs. This way duplicate-heavy use cases...

Thanks @evgenydmitriev . Looks like the most expensive part is Vectorize calls. I'm intending to add [Cloudflare cache API](https://developers.cloudflare.com/workers/examples/cache-api/) over each individual vector. Price given 1000 requests: **Workers AI** https://developers.cloudflare.com/workers-ai/platform/pricing/...

Actually, you're right @evgenydmitriev ! Obviously, a database is a better and more straightforward solution than distributed cache here, so _now_ I'm planning to use a [D1](https://developers.cloudflare.com/d1/examples/d1-and-hono/) to create a...

Hi @evgenydmitriev ! I have some updates to bring up to you. Firstly, [worker limits](https://developers.cloudflare.com/workers/platform/limits/#account-plan-limits): - max 6 simultaneous outgoing fetch requests - 128 MB memory - Duration - no...

@evgenydmitriev, It would be nice to have URL like this - GET https://www.example.com/some-namespace?text=some%20text, and then let Cloudflare CDN cache those. I've tried this, and it doesn't work because workers are...

@evgenydmitriev, well, latest suggestion is to start small and create an endpoint that works on a single message, but is cacheable by URL. Then either have clients initiate multiple parallel...

Now it's a POST body, so shouldn't be cacheable by default. In this prototype, I converted to GET (with no luck admittedly 😀)

Hi @evgenydmitriev , please review updates.

Hey @svencoeck , thanks for your contribution! Can you please just polish a little: - we try to indicate roughly applicable Magento version in file name, e.g. `Patch-Magento_Framework-M2.2.3-fix-13929-images-cant-be-uploaded-in-symlinked-folders.patch` - add...