Trieve

Today we are launching PDF2MD! It's the easiest way to convert documents with complex layouts into LLM-ready markdown. Infographics, tables, patents, litigation filings, and more all become possible when LLMs with human-level intelligence get involved. This complex Nvidia revenue flowchart was converted for just $0.00400. Try it yourself, link in comments ?? .

We are going to be releasing our own take on using the vision LLM's to convert PDF's into markdown on Thursday. Here's a preview of a cooler conversion result. Mind blowing!

WE ARE HIRING SOMEONE TO WORK ON OPENAPI TOOLING AT TRIEVE!! We have a large 16 resource, 70-route API and it’s hard to build high quality SDK’s. Existing OpenAPI generator solutions are end-to-end. They do types, docs, and request handler functions. We need a generator that only does types + boilerplate request handler functions and lets us fill-in-the-blanks for those functions. Downstream, it needs to insert the handler functions into the OpenAPI spec as x-codeSamples. We are calling this a “headless OpenAPI generator”. Do you resonate with this? Do you want to help us build a successor to the openapi-generator project? If so, apply below or DM! https://lnkd.in/gJzRw49y

Today, we're making Sitesearch available to all - for free ?? ? Fact: 62% of programmers who use AI tools use them to search for answers. Unlike traditional docsearch, Trieve understands your code, technical content, and users' complex queries without sacrificing relevance, speed, or control. Meet your users' demands with a simple drop-in component. It's an easy win! To celebrate, we're launching on Product Hunt. We'd love to have your support! https://lnkd.in/e7SEn2BU

??? 2 days until we launch on Product Hunt

We have released verion 0.12.0!! The high level features are: - Better Analytics, as a first-class member in the dashboard - New Scraping support - Search and Rag components See the full release ??

NEW BLOG! - ? PGVector's Missing Features ? We use PostgreSQL along with a dedicated search solution (Qdrant) and our own server which is complex and not ideal. It would be great if we could simplify and use PostgreSQL by itself, but, outside of scale, it's missing features which prevent us from making the swap. We share what those are in this blog! https://lnkd.in/gvKAak26

Wooohoo! ??

For Launch Week Day 2, we are announcing our new JS/TS SDK for building RAG and AI Search ? It's easier than ever to build amazing AI applications using Trieve ? See the full blog ?? https://lnkd.in/gS8iFEVp

To kick off Launch Week, we're bringing typo tolerance to semantic search Vectors aren't a cure-all and fail to understand typos by themselves Today, we're bringing blazingly fast spellcheck to retrieval pipelines everywhere ??