Recommendations for CPU-Based Real-Time Vector Database Indexing and Matching?

Hello everyone, I have a specific online vectorization use case: I’m looking to search the internet for articles, vectorize these articles along with the search queries, and then retrieve the most relevant passages from them. Currently, I have basic hosting through DigitalOcean.

Could anyone recommend the most suitable vector dataset for this task? Additionally, considering my resources, is it feasible to run this system solely on CPUs? And if so, would this setup be scalable if deployed on CPUs only?

And you don’t want to use OpenAI embedding products?