The Alpha Signup of Advanced AI Software Made Simple

We are thrilled to announce the Alpha release of Prism, a full-service platform for Retrieval Augmented Generation (RAG) and AI development. Prism offers an enterprise-grade API, high-bandwidth interface for information retrieval (including integration with all your favorite cloud platforms), as well as a user-friendly front-end knowledge-base editor, enabling you to effortlessly harness the full potential of AI technology, without advanced knowledge in Data Science or Machine Learning. In the next 4 weeks, we are launching the most advanced RAG API on the market to date, allowing you to fully integrate your text files and information with Large Language Models (LLMs).

In celebration of this milestone, we are extending an exclusive invitation to the first 1’000 users to sign up for our platform, with a discount on the first year, and 10'000 free API tokens!

AI Software Development Made Easy

At Prism, we understand the challenges involved in building AI software. That’s why we have simplified the process to ensure a seamless experience for our users. With our out-of-the-box RAG (Retrieval Augmented Generation) solution, connected to a powerful API, and supported by libraries in various languages (Python and Javascript), you can now supercharge your AI language model in just a few clicks or a few lines of code.

import prism_ai as pai

kb = pai.KnowledgeBase.create( 
  name = 'My Files', 
  base_dir = '/path/to/my/files/' 
)      

reply = prism_ai.Reply.stream( 
  prompt = 'tell me about my files', 
  kb_id = [kb.id]
)

Turn your Cloud Storage into a Knowledge Base

Retrieval Augmented Generation (RAG) is a technique that enables you to retrieve information from knowledge sources in order to enhance the capabilities of large language models. By seamlessly integrating your own files, websites, or cloud storage, Prism allows you to create custom language models tailored to your specific needs. Whether you envision building an AI legal expert, a research assistant with up-to-date knowledge, or a user-friendly chatbot for your website, Prism’s RAG API empowers you to quickly integrate your textual information, and rapidly construct high-performing systems.

Exciting news! We’ve just released a demo showcasing the incredible capabilities of our product. Curious about how to build your AI app in minutes with Prism? Check out this video:

Fully Private, Fully Transparent

We take Data Privacy and Security very seriously. We understand that you don’t want your data shared with external firms or AI systems, which is why we are offering our enterprise platform, complete with fully private instances of our software. That way, you can run Prism in a fully contained environment, with no external communication to Prism, OpenAI, or the like. Your servers, your cloud, your terms.

More than Just a Vector Database

While traditional techniques for RAG rely solely on linear search over databases of vector embeddings, Prism introduces several improvements which aid greatly in the performance of our RAG system:

• Multidimensional Index Structures. While linear search is always guaranteed to find the most relevant context (in terms of embedding similarity), this method degenerates over time, as query times scale linearly with the size of your database. We introduce our multidimensional index structure (Smart Index), in two flavors: Heuristic and Precision indices, which nearly always provides the best context, and always provide the best context, respectively. Both algorithms offer a significant improvement over linear search and provide means of performing RAG over massive sets of textual context.
• Named Entity Retrieval. While vector embeddings excel at capturing similarities between phrases, they struggle to capture relationships between named entities or complex concepts. This hinders their ability to generate contextually appropriate responses and limits their understanding of specific domains. Prism mitigates this shortcoming, native to standard vector retrieval, by extracting and retrieving named entity context and relationships into our Knowledge Bases.
• Hyde. HyDE is an embedding technique that takes queries, generates a hypothetical answer, and then embeds that generated document and uses that as the final example. This technique has been used, with great success, to improve the context provided for generations based on very short user prompts. Prism ships with Hyde built in, so you can execute it by changing a single parameter.
• Scalable Precision. Standard RAG techniques treat all context as equal, by selecting a fixed window size, and storing all raw text as a single size. We provide via our simple API interface, a simple way to dynamically change the window size of raw text to be stored. This way, the important stuff gets retrieved when it needs to be.

Join Prism’s Alpha Release and Sign Up Now.

We are excited to invite the first 1’000 users to join us on this groundbreaking journey. As our valued early adopters, you will be the first informed of all new updates, upgrades, and of course, receive some sweet discounts! This is your opportunity to experience the simplicity and power of building advanced AI software without any financial commitment.

Prism’s Alpha release marks a significant advancement in democratizing AI software development. By offering an enterprise-grade API, a user-friendly knowledge-base editor, and the ability to tap into Retrieval Augmented Generation (RAG), we have empowered developers and users to unlock the true potential of AI technology. Sign up today to be one of the first 1’000 users and embark on a transformative journey in AI innovation with Prism. Together, let’s shape the future of AI software development.