Retrieval SDK
The Pryon Retrieval Engine is an end-to-end pipeline that allows you to ask questions about your data. It takes in a natural language query, and returns a response, along with reference context retrieved which can then be passed to the LLM.
Retrieval returns the following outputs (scored and ranked):
-
Answer snippets
-
Corresponding context chunk or answer in context (AIC)
-
Document image previews with highlighted answers
-
Links to source documents for attribution
By using our retrieval engine as a key, foundational component of the RAG stack, organizations can quickly build generative AI-powered experiences for customers and employees.
Generative SDK
The Pryon Generative SDK offers customers a retrieval augmented generation (RAG) experience that utilizes the Pryon Ingestion Engine for unstructured data ingestion and the Pryon Retrieval Engine for question answering.
The Generative SDK gives you access to:
- G-LLM orchestration and prompt engineering: Flexibility to choose a public, custom, or Pryon-developed LLM.
- Model evaluation: Choice of which LLM that best meets your needs based on various measures, such as speed and accuracy.
- User feedback: Ability to collect feedback by rating responses with a simple and intuitive thumbs up/thumbs down reaction.
The Pryon Generative Engine offers a safe and trustworthy pathway to implement enterprise-grade generative AI, prompting the LLMs to only produce output based on an organization’s trusted content, which mitigates the security and hallucination risks of public generative tools.
Who can use the SDK
While the Ingestion SDK is available to all signed-in users, only Enterprise users currently have access to the Retrieval and Generative SDKs. To become an Enterprise user, schedule a demonstration.