Introducing Cyndalf

It’s become increasingly evident that AI agents are the future of human-machine interaction. Mid-size companies deploy dozens of security solutions. Large companies often have over a 100. Security is a landscape of point solutions. Buying security has always relied on some form of tribal knowledge. That friend or private slack group you trust is your go-to resource when you are deciding which MDR or integration solution you want to go with. You often want to complement community wisdom with evidence and data. But navigating hundreds of options to shortlist a few and understand them deeply takes effort. Cyndalf makes this effortless. Engineered with Retrieval-Augmented Generation (RAG), Cyndalf is trained on millions of data points on an evolving vendor landscape and external news that impacts vendor choice. Unlike a GPT wrapper, it is built to evaluate thousands of vendors and their products with information retrieval based on a custom, secure set of millions of data points.

Share This Post

Technical Overview

Cyndalf was designed to leverage the sophisticated capabilities of a RAG, using the data retrieval capacities of OpenSearch and the nuanced generative abilities of advanced AI models like GPT. This section provides a technical deep dive into how RAG operates within Cyndalf and why it was chosen as the foundational technology.


RAG: Bridging Retrieval and Generation

The model first queries an extensive database—in Cyndalf’s case, a specialized OpenSearch database containing detailed data points on cybersecurity vendors, products, ratings, and feature sets. It then utilizes this retrieved information as context to generate informed, precise responses.

The RAG process involves several key steps:

  1. Query Processing: Cyndalf processes the user’s cybersecurity inquiry, breaking it down into manageable queries that can be effectively addressed by the database.
  2. Information Retrieval: The processed queries are then matched against the OpenSearch database, retrieving the most relevant documents and data points related to cybersecurity solutions, vendors, and products.
  3. Contextual Synthesis: Leveraging the retrieved information, RAG employs a generative AI model to synthesize responses that are not only accurate but also tailored to the specific context of the inquiry, ensuring relevance and depth.


Why RAG Over OpenAI Custom GPTs?

While OpenAI’s custom GPT models offer broad generative capabilities, their generalized nature and training on diverse internet datasets make them less suited for specialized applications like Cyndalf. RAG, by contrast, provides a focused approach, ensuring that the generative component of the model is always informed by the most relevant and up-to-date information from the OpenSearch database. This makes a RAG more robust when increasing the variables to make informed vendor choices. The latest news and threat intelligence for example, can begin to layer on top of this foundation. 


Implementing RAG in Cyndalf: Use Cases and Advantages

The implementation of RAG in Cyndalf enables a range of functionalities tailored to the diverse needs of cybersecurity professionals. This section explores specific use cases and the inherent advantages of Cyndalf’s RAG-based approach.


Use Cases

  • Complex Query Resolution: Cyndalf can dissect and address complex queries that require a deep understanding of cybersecurity concepts, leveraging RAG to provide comprehensive, contextually informed responses.
  • Real-Time Threat Analysis: In response to inquiries about emerging threats, Cyndalf utilizes real-time data retrieval to provide up-to-date information and recommendations, a capability made possible by the dynamic nature of RAG.
  • Customized Solution Generation: For users seeking tailored cybersecurity solutions, Cyndalf employs RAG to synthesize bespoke recommendations, drawing on a wealth of data regarding vendors, products, and their features.



  • Precision and Relevance: By integrating RAG, Cyndalf ensures that every response is not only generated with high accuracy but is also deeply relevant to the user’s specific query, a critical factor in the complex field of cybersecurity.
  • Data Security and Integrity: Cyndalf’s reliance on a secure, internal database for retrieval ensures protection from data poisoning.
  • Scalability and Efficiency: A RAG allows Cyndalf to efficiently handle an increasing volume and variety of queries, showcasing  scalability without compromising response quality.


Cyndalf is a co-pilot custom-built for cybersecurity practitioners. It is extensible for buyers, consultants, researchers, investors, marketers and CISOs who want a more intuitive way to make sense of complex provider choices. Integrating advanced retrieval and generative capabilities, Cyndalf is an answer to Gandalf’s (of Lord of the Rings fame) timeless quote: “ All we have to do is decide what to do with the time that is given to us


Watch Cyndalf in action

Play Video about Cyndalf

More To Explore


Introducing Cyndalf

It’s become increasingly evident that AI agents are the future of human-machine interaction. Mid-size companies deploy dozens of security solutions. Large companies often have over a 100. Security is a

Cypher Score

One Score To Tie Them All

Buying a cybersecurity product or service has costs- monetary and otherwise. Picking the wrong provider, even more so. Purchasing decisions that concern business security are rarely made without significant due

By visiting you accept our use of cookies and agree to our privacy policy.