Introduction

What is Trellis?

The Trellis engine ingests any unstructured data sources and converts it into structured SQL format based on the schema and transformation you define in natural language.

Typically, processing and analyzing unstructured data like PDF and html files are highly manual. It takes many engineering hours to build a scalable data pipeline, parsing engine and deploy a cost effective model.

⛳️ Trellis allows you to:

  • Enrich your database with business critical data
  • Go beyond RAG and run complex SQL queries (like aggregation and trend analysis) on top of your unstructured data.
  • Quickly iterate on new features without spending time building out complex ML pipeline.
  • Add new AI features in minutes without the need to handle messy data
  • Monitor data quality and flag anomalies before it goes into production

✅ Some key features include:

  • Define the LLM operations and data schema directly in YAML. No extra configuration.
  • Automatically combine LLMs and rule-based heuristics to deliver the best performances.
  • Horizontally scaling infrastructure based on data volume. No need to manage your own GPU and CPU cluster for data processing.
  • Robust tracking of different experiments status

Discover Trellis

Quickly get started with Trellis through our example guide:

Learn more about Trellis platform and how it works here

Access the API docs here