Show HN: In-memory Documents Search Engine for structured and unstructured data

[sagar-co]

Hi HN, We are excited to launch an in-memory documents search engine. Please check out the demo at https://finance-demo.joyspace.ai.

With this demo, we indexed around 500,000 pages of Mutual Fund Prospectus Risk/Return Summary Data Sets from the SEC. This includes mutual fund prospectus and other disclosures by mutual funds.This form is called 485BPOS document (and 497 form). We indexed a couple of quarters worth of data but we can easily scale to 20 years of data and still serve results in milliseconds. You can extend this to 10K, 10Q, 8K, and hundreds of other finance documents.

For this demo, we built UI showcasing search capabilities. You can use pagination, facets and filtering, search-as-you-type, grouping out of the box. It is possible to integrate with top UI frameworks such as ReactJS, SvelteJS, VueJS, or plain Javascript.

You can also extend this capability out-of-the-box for legal and other industries. You can also power your own AI assistant with our search.

We have APIs for indexing and search. We support all standard formats such as JSON, PDF, CSV, TSV, HTML, DOC, DOCX, RTF, and others. We manage indexing and search pipelines for you.

Let us know what you think.

Video of our demo site: https://www.youtube.com/watch?v=l3Rbhsp1eUg

[sagar-co]

Hi there, I am a co-founder of Joyspace.ai. If you have any questions/comments, feel free to contact me.