HomeGuidesRecipesAPI EndpointsRelease NotesCommunity
Log In

Available API Classes

API Classes Overview

We've broken down the different API classes into a few categories for easier referencing. Within each category will be a list of available API classes, each with their own methods and respective documentation.

Data Fabrics

This collection of API classes allows you to create and manage your data fabric: what connectors go into which instances, which instances are a part of a specific data fabric, and more.

dataFabric – Search for data fabrics, or the Layar instances within a given data fabric.

connector – Create or maintain real-time streaming connectors such as Twitter or RSS feeds.

liveSource – Create or maintain the scheduled index jobs of connectors such as S3 buckets, HANA databases, and Google Drives. Note: this endpoint does not capture or maintain local file uploads.

savedList – Create, add, or retrieve items from a saved Layar Set of documents, images, or other types of data.

Layar Search

This group of API classes allow you to search your data fabric for Cortex Domain Objects like documents, paragraphs, statements, and more.

all – Search across all existing Cortex Domain Objects within a given Layar instance for a specific record.

autocomplete - Retrieve your search history.

sourceDocument – Enable a whole bunch of capabilities around the retrieval and utilization of documents pulled into Layar. Search, upload, annotate, download raw files, extract images or tables, and much more - there's so much you can do! This endpoint has a full walkthrough of features, which can be viewed here.

paragraph - As part of our automated document processing, we parse out the different paragraphs of your documents. You can retrieve and manage those paragraphs here.

statement – We also parse out statements (snippets of text that comprise a paragraph), which we use for many question answering and named entity recognition tasks. You can retrieve and manage those statements here.


Table Contents Are A Collection of Statements

It's worth noting that each row in a table is considered its own statement object. For example, a table with 100 rows of clinical trials and 20 columns of properties would still only have 100 statements.

Question Answering (QA)

These classes allow users to ask natural language questions and retrieve answers from content within a data fabric, using our QA Service. You will often see these endpoints used in both open- and closed-domain question answering tasks.

question - Create and manage all the questions you wish to ask of the QA Service.

answer - Search and manage all the answers derived for the questions asked.

Named Entity Recognition

This collection of APIs allow you to tag terms and phrases from text as categorical concepts using named entity recognition. Tag text using your own custom named entity models or Vyasa-provided ones. Learn more!

namedEntity – Search and manage a list of currently available NER models.

concept – Create, manage, and delete named entities (aka “concepts”). Identify, assign, and extract hierarchical relationships between terms and their named entities (i.e. preferred synonyms).

conceptType – Search and manage concept types (categorical groups such as 'gene' or 'disease').


ontologyTerm – Manage ontologies using these endpoints. You can create, edit, search, and manage ontologies being used within your instance.

Layar Relations

A collection of classes of text analytics methods that will help users with pre- or post-processing and data cleanup tasks.

group - Bundle like-terms (“Vyasa Analytics”, “Vyasa Analytics, LLC” ) together into groups based on a fuzzy string match. into a single grouping or single output term.

radar – Identify similar words and phrases that exist within close proximity in vector space.

vectorSimilarity – Identify similar statements that exist within close proximity in vector space.

Compound Analytics

Run a toxicity prediction model or render chemical structures using these endpoints.

compound – Render SMILE strings into compound images.

smiles – Run toxicity models and create beautiful visualizations for an input of compound strings.


Your data scientists and MLOps teams can use these classes to manage and modify the models used by Layar for projects with analytics pipelines (e.g. leveraging new NER models, fine-tuning and deploying a custom QA model, etc).

model – Search and implement Vyasa’s commercially available pre-built models to assign to different modules in a project. Users can also upload their own custom models with this API.

module – Within a project, users can define the type of module they are performing (ie image classification). Search or save modules implemented by Vyasa for an assortment of use cases to use in custom projects for similar applications, including fine-tuning a transfer model for a private dataset.

project – Create, edit, or remove projects from a Layar instance. Projects have a versatile design that may include several results as a component of the project management endpoints.

projectComputation – Output files for any quantitative readouts implemented by the user as defined by the project and module. Most users utilize this API for output files pertaining to number of iterations, model assessments, etc.


Your DevOps and SecOps teams can use these classes to pull logs on performance, availability, and data access records.

status - View the status of your Layar instance (whether it is running or down).

event – View metadata for an event performed with a Layar API, such as user, date of event logged, project ids, and more.