Idyl E3 on the AWS Marketplace Quick Start

After launching an instance of Idyl E3 through the AWS Marketplace, the instance and Idyl E3 will start. After a couple of minutes, Idyl E3’s API will be available at http://idyl_e3_ip_address:9000/api, where idyl_e3_ip_address is the IP address or DNS host name of the instance. Idyl E3 is now ready for entity extraction requests.

Idyl E3 includes an English-persons entity model. This model is intended for test and evaluation purposes because it will probably not give decent results in production use. However, it is perfect for learning how Idyl E3 works. Let’s submit an entity extraction request to our new Idyl E3 instance.

Submitting an Entity Extraction Request

To submit an entity extraction request we are going to use cURL. The cURL command is:

curl -X POST http://idyl_e3_ip_address:9000/api/extract?text=George+Washington+was+president.

This request sends the text “George Washington was president.” to Idyl E3 for entity extraction. (Be sure to replace idyl_e3_ip_address with the IP address or DNS host name of your Idyl E3 instance.)

Idyl E3 will process the request and return to you the following:

  "entities": [
      "text": "George Washington",
      "confidence": 0.89,
      "span": "[0..2)",
      "type": "person",
      "enrichments": {
      "languageCode": "en",
      "context": "not-set",
      "documentId": "not-set",
      "extractionDate": 1471465494972
  "extractionTime": 1

The response tells us a few things. First, we see that Idyl E3 identified “George Washington” as a person entity having a confidence of 0.89. Confidence values range from 0 to 1, with higher values indicating a higher probability of having an actual entity.

The span indicates the location of the entity phrase “George Washington” in the input text. The span [0..2) means that the phrase “George Washington” was the first two words in the input text.

The type tells us that “George Washington” was identified as a person entity.

The enrichments are empty because Idyl E3 out of the box is not configured to perform any entity disambiguation or entity enrichment.

Since we did not provide a language in our entity extraction request, Idyl E3 defaults to executing all entity models for all languages. The languageCode tells us that the “George Washington” entity was identified by a English-language entity model.

The context and documentId values allow us to identify the sources of text. We did not provide a context or a documentId in our entity extraction request so they both default to “not-set.” (You can think of a context as a book chapter and a documentId as a page in the book. They are both optional in entity extraction requests.)

Lastly, the extractionDate is a Unix timestamp of when the extraction happened.

EC2 Security Groups

The AWS Marketplace launch wizard will create an EC2 security group that permits inbound connections on ports 9000 (the Idyl E3 API), 22 (SSH to the host EC2 instance), and 80 (for the Idyl E3 documentation).


When launching Idyl E3 via the AWS Marketplace you will have a fully functional instance of Idyl E3 in just a few minutes. When ready, Idyl E3’s API will be ready to process entity extraction requests. See the full documentation of the Idyl E3 API. You can certainly use cURL to send all of your entity extraction requests to Idyl E3 but you may find our Idyl E3 client SDKs useful. The client SDKs are available on our GitHub.