Getting Started with Squirro#

Squirro is a software platform that helps you connect many different types of data sources together, including unstructured data like PDFs, Word and Excel documents, call recordings, and more.

Squirro can be installed locally, or in the cloud, with end users interfacing with the platform via a web browser.

Once your data is connected, Squirro helps you:

  • enrich, connect, and understand your data using Artificial Intelligence techniques.

  • visualize your data in intuitive ways.

  • search across all of your data.

  • gain actionable insights into your data.

This page introduces the Squirro platform and highlights they key parts of installing Squirro, connecting data sources, processing and visualizing your data, then searching and gaining insights from it.

Use Cases#

Almost any organization can benefit from Squirro, but Squirro is particularly useful for enterprise organizations with large amounts of siloed data, or organizations that need to connect data from many different sources together.

To learn more about specific solutions, click any of the following to bring up its associated case study:

Knowledge Management Insights

Cognitive Search

Knowledge Management Insights

Cognitive Search

Service Insights

Machine Learning Classification

Understanding Squirro#

The Squirro platform can be understood in terms of a Gather, Understand, Act framework.

The following diagram illustrates how data is connected to the platform and is processed through it, before being visualized and acted upon at the end (working from left to right):

Ingester Service and Data Processing Pipeline

An important part of this is the Data Processing Pipeline, which is where, through AI and machine learning techniques, data is enriched, connected, related, classified, predicted, and understood.



Squirro is built with security in mind from the ground up.

As an ISO 27001-certified company, Squirro is committed to security and privacy, with a strong emphasis on protecting client data.

Learn more about Security at Squirro.

Installing Squirro#

There are three ways to install Squirro:

  1. Install Squirro as a Service.

  2. Install Squirro on a local server.

  3. Install Squirro on a private cloud.

Squirro as a Service#

Squirro as a Service is the easiest way to get started with Squirro. It is a fully managed Squirro installation, hosted in the cloud by Squirro, and available to you 24/7.

To get started with Squirro as a Service, follow the steps below:

  1. Visit

  2. Register for a Squirro ID.

  3. Install either a full Squirro application or try an online demo, as shown in the screenshot below:

Squirro Self Service page

Install Squirro Locally or on a Private Cloud#

If you prefer to install Squirro locally or on a private cloud, you can do so using Linux.

While you can manually install Squirro on Linux either online or offline (see Installing Squirro on Linux), Squirro recommends using Ansible to install.

Reference: See Install and Manage Squirro with Ansible for step-by-step instructions.

Connecting Data Sources#

Quality in, quality out.

Squirro’s search results, data visualizations, and insights are only as good as the data you connect to the platform.

After installing Squirro, the first thing you’ll do is connect your data sources.

Although Squirro has a Command Line Data Loader tool, the easiest way to connect data sources is to use the UI Data Loader, as shown in the screenshot below:

Example Data Loader

Processing Data#

Squirro’s data processing can be understood as a series of steps, or pipeline steps, that data goes through as it is processed.

The Pipeline#

The Squirro Pipeline (referred to as simply the Pipeline or pipeline) is where connected data undergoes a series of transformative steps to become useful, actionable data.

Reference: Learn more about the pipeline generally at Data Processing Pipeline.

You can manage every step of the pipeline within the Squirro UI using the Pipeline Editor, as shown below:

Squirro Pipeline Editor

Built-in pipeline steps include the following:

  • Enrich Pipeline Steps, used to enriche data through actions like unshortening links, duplicate detection, PDF OCR, language detection, noise removal, and more.

  • Relate Step, which connects data together based on relationships you define using Squirro’s proprietary Known Entity Extraction.

  • Discover Step, powered by the NLP tagger for key-phrase extraction, named entity recognition, and rule-based sentiment analysis.

  • Classify Steps, which use machine learning models created in AI Studio to classify data.

  • Predict Step, generated from Trend Detection.

  • Automate Pipeline Step, allowing for the automation of tasks like sending emails.

  • Index Steps, needed to persist Squirro items on disk for search and visualization. This includes content standardization, cache cleaning, indexing, and search tagging and alerting.

  • Flow Step, which is used to manipulate the direction of the items in the pipeline workflows.

Machine Learning#

An important part of the data processing pipeline is machine learning.

Squirro uses machine learning in three contexts, as part of its:

  • libNLP, Squirro’s Natural Language Processing library, which powers AI Studio, query processing, and other parts of the platform.

  • Model-as-a-Service, which allows you to import custom machine-learning models into Squirro.

  • AI Studio, a no-code tool for creating machine learning models you can use to classify data from your browser with a five-step process, as shown in the screenshot below:

Proximity-Rule Quality Explanation

Visualizing Data#

Squirro’s data visualizations are designed to help you understand your data and to help you gain insights into it.

Squirro uses Dashboards to deliver user experiences for search and insights.

Applications come with dashboards pre-configured, but you can extend those dashboards or create your own, with almost no limit to what you can do. This is thanks to Squirro’s drag and drop Dashboard Editor, shown in the screenshot below:

Squirro Dashboard Editor

Reference: See Dashboards for a full list of Dashboards documentation.

Layers and Widgets#

Squirro dashboards are made up of Dashboard Layers, each containing one or more Widgets.

Layers are used to group widgets together, and to control how they are displayed using Visibility Conditions.

Built-In and Custom Widgets#

Out of the box, Squirro comes with many built-in widgets that allow you to visualize, search, and interact with your project data.

Reference: See Built-In Widgets for a full list of built-in widgets.

You can also extend Squirro dashboards by creating your own custom widgets.

Reference: See Custom Widgets to learn more.


Communities are a Squirro feature that allows users to personalize their experience by following topics of interest to them, visualized as shown in the screenshot below:

Squirro Communities

Individual communities are grouped together into Community Types that are defined by the Project Creator.

For example, as shown in the previous example screenshot, a Project Creator might define a community type called Food Products and then create communities for each of the foods represented within the project, such as Bakery Products, Cocoa Coffee and Tea, Dietetic Foods and so on.

They are powerful means of allowing users to identify subject areas of interest and receive tailored, relevant content.

Reference: Learn more about Communities.


Insights aren’t a specific out-of-the-box feature of Squirro, but rather a combination of Squirro’s data processing and visualization capabilities.

The possibilities for generating and visualizing insights using Squirro are almost limitless.

Example Use Case#

In the Business Environment Monitoring (BEM) use case, Squirro is used to monitor the business environment of a company, including news, social media, and other sources.

Reference: See Business Environment Monitoring Quick-Start Guide to try it out yourself.

Using machine learning models, Squirro can identify and classify the topics of interest to the company, and then visualize those topics in a dashboard.

In the example screenshot below, those insights are visualized as Chips that fall into one of the identified BEM insight categories:

BEM Guide Example

End users can click on a chip of interest to bring up a list of documents that are relevant to that chip.

Clicking on the Bankruptcy chip, for example, brings up a list of documents that contain information about bankruptcies. Opening a document will then

Troubleshooting and Support#

If, at any point along the way, you run into trouble, you can ask a question in the Squirro Forum or contact the friendly folks at Squirro Support.