Skip to main content

Extract and generate data

In this guide, you'll learn how to use LeMUR to extract data such as tags and descriptions from your audio.

tip

If you want a Quickstart, see Apply LLMs to audio files.

Before you start

To use LeMUR, you need an with a credit card set up.

With the LeMUR Task, you can send any prompt to the LLM and apply the model to your transcribed audio files.

To extract data from your audio, define a prompt with your instructions and call transcript.lemur.task(). The underlying transcript is automatically used as additional context for the model.

Example output​

Runner's knee
Injury: Runner's knee
Symptoms: Pain behind or around the kneecap, Pain when walking

Sprained ankle
Injury: Sprained ankle
Symptoms: Pain and difficulty moving the ankle, Swelling around the ankle, Bruising

Meniscus tear
Injury: Meniscus tear
Symptoms: Stiffness and swelling, Pain in your knee, Catching or locking of your knee

Rotator cuff tear
Injury: Rotator cuff tear
Symptoms: Pain when lifting and lowering your arm, Weakness when lifting or rotating your arm, Pain when lying on the affected shoulder

ACL tear
Injury: ACL tear
Symptoms: Severe pain and tenderness in knee, Loss of full range of motion, Swelling around the knee

Data extraction prompt examples​

Try any of these prompts to get started:

Generate titles and descriptions

Generate metadata information about your audio, such as title and description.		"Generate an attention-grabbing YouTube title based on the video transcript"

Generate tags

Generate tags to organize and categorize your audio data.		"Generate keywords that can be used to describe the key themes of the conversation"

Action items

Extract action items from the meeting transcript and assign them to the corresponding speaker.		"What action items were assigned to each participant?"

For more use cases and prompt examples, see LeMUR examples.

Improve the results​

To improve the results, see the following resources:

© AssemblyAI