Subscribe
Sign in
Vectorize ā Build RAG pipelines that are optimized for your data.
Vectorize is a data platform for retrieval augmented generation (RAG). It combines RAG evaluation to identify the best way to vectorize your data with a cloud-scale RAG pipeline engine. Vectorize populates your vector database and keeps your vector data fresh.
27
Replies
Best
I'm an idiot when it comes to this stuff, but it seems this would be great for e.g. generating content in the style of e.g. my twitter history, or like, articles on Techcrunch?
@jonnymiles If you have a bunch of samples of the writing style you want to see it can definitely help, but one of the things pre-trained are worst at is matching writing styles in my experience.
Just signed up and scheduled a call for next week! Looking forward to giving this a spin!
Great work @chrislatimer !
@alexo_125 You can either query your vector database directly or we offer an API endpoint to retrieve data that includes some built in features to help simplify your RAG application. The API provides built-in reranking and will vectorize your input query for you so you don't have to worry about those things in your app.
This is extremely useful for anyone building LLM applications. Currently evaluating!
Vectorize was a game changer for us at FamilyCloud.AI. We are building FamilyCloud to help families get organised effortlessly.
FamilyCloud was initially conceived as a SaaS product but with all the development in LLMs, it made sense for us to (re)conceive this as an AI native application.
Vectorize helped us get our private beta out in record time (in a matter of weeks, to be more specific they help us set up our RAG pipelines in a matter of days).
Now FamilyCloud members (families) can upload data onto our proprietary storage and ask questions of that data and retrieve information and insights.
Vectorize solves the problem of connecting to varied data sources , extracting the content, converting them (text, images etc.) into these mathematical representations called vectors (a prerequisite for all RAG applications) and storing them into vector databases.
They also optimise the chunk length automatically so as to get the most optimal response.
What's more - they also ensure that the vectorisation happens immediately after upload so that no stale data is served to customers.
If you are building a scalable RAG application, I would strongly recommend using Vectorize.io. The founders, Chris L and Chris B have been great partners for us at FamilyCloud.AI
@chrislatimer
Congrats on launching Vectorize! Excited about how it helps with RAG and keeps data fresh. The cloud-scale RAG pipeline engine sounds powerful. Great for anyone needing efficient data vectorization.
Love your concept @chrislatimer. A cloud scale RAG pipeline engine combined with automated vectorization sounds like a powerhouse for data handling. Well done.
Huge congrats to the Vectorize team on today's launch! I'm intrigued by the promise of optimized RAG pipelines tailored to specific data. Quick question: How do you handle scenarios where the underlying data structures or schemas are constantly evolving - does Vectorize's auto-vectorization adapt to these changes in real-time?
@ihuzaifashoukat We primarily work with unstructured data so the data we're ingesting so there's not usually a schema involved.
@chrislatimer does Vectorize support hybrid search, like combining both traditional keyword and vector search? Overall good work.
Hi @ava_bailey0, currently it's just vector, but we are partnered with Elastic and have been talking to them about supporting BM25 and possible ingestion options there. What would you like to see?
I recently tried Vectorize, and I have to say, itās quickly become one of my top RAG tools. Unlike other similar tools that can be slow to run, Vectorize impressed me with its speed and accuracy, delivering truly mind-blowing results. Huge thanks to Chris and the team for creating such an outstanding tool.
Vectorize