Enter the attribute name and value below. This attribute will be applied to {numFiles, plural, =1 {your 1 selected file} other {each of the your # selected files}}.

Name

Use the following CSV template to add attributes to {numFiles, plural, =1 {your 1 selected file} other {each of the your # selected files}}. When you are done, upload the completed file.

Upload your image, video, or audio files that you want to use to analyze or train expression models

Account security

The Hume AI Platform can be integrated into any application or research study that involves human data: audio, video, images, or text. Use our APIs to access models that measure upwards of 50 dimensions of emotional expression in nuanced facial and vocal behavior. Capture subtle expressions like boredom and desire in the face, vocal expressions like sighs and laughs, continuous emotional intonation in speech, emotions conveyed in text, and moment-to-moment multimodal estimates of emotional experience.

Create your Hume AI account

You can use these accounts to log in.

Connected accounts

We can’t find an account associated with this email. Please sign up.

An email from platform@hume.ai has been sent to {email}. Please click on the link inside to verify your account. You may need to check your spam folder.

Please verify your email

Please check your inbox for an email from Hume to continue with our Platform

Verify your email

Notifications

There are a couple ways you can use this model

Sign up to train a custom model aligned to your specific needs

Create your own Custom Model

Add files you have previously uploaded to Hume

Please confirm that you want to replace your entire dataset with the contents of your CSV file.

Please remove these files and re-upload them in order to use this dataset to create a custom model.

Untitled Dataset

Upload individual image, video, audio, or text files.

Upload a folder containing image, video, audio, or text files. We'll use the folder name as the label for all files inside.

Dataset columns

Data consistency

Data quantity

Files are valid

Create your first Dataset

Please provide at least 2 labels. Then, upload files for each label.

We’re experiencing longer than usual processing times as we scale. The team is actively investigating this. Thank you for your patience.

High request volume

We encountered the following issues while analyzing your files.

{numErrors, plural, =0 {} =1 {1 Error} other {# Errors}} Encountered

We encountered the following issues while processing your job.

Differentiate 37 kinds of facial movement that are recognized as conveying distinct meanings, and the many ways they are blended together

Discover over 25 patterns of tune, rhythm, and timbre that imbue everyday speech with complex, blended meanings

Explore vocal utterances by inferring probabilities of 67 descriptors, like 'laugh', 'sigh', 'shriek', 'oh', 'ahh', 'mhm', and more

Leverage our expression models to train a custom model aligned to your specific needs

Create your first Custom Model

"<strong1>{selectedTargetFeature}</strong1>" is a {featureType} column. As such, we recommend <strong2>{defaultTaskType}</strong2> as the task type for your model.

Your model is now training. You can check on its progress in the Jobs section.

All done!

For a successful classification training run, your target column will need at least {minNumLabels, plural, =1 {1 label} other {# distinct labels}} and {minNumFiles, plural, =1 {1 or more file} other {# or more files}} per label.

We have detected an issue with your labels

Train your first model

Extract hundreds of expression measures from any saved file

Upload multimedia files

On our Playground, explore the nuances of emotional expression in video, audio, images, and text

Welcome to Hume

Uncover emotions, sentiment, and toxicity conveyed as you type

Live text analysis

Get rapid, real-time expression measures for streaming data

Livestream through webcam

Use custom models that understand expression and language

Interpret expressive communication

Read the Hume Initiative guidelines for creating empathic AI

Ethical guidelines

Use psychologically valid models of facial movement and vocal modulation

Expression Measurement

Learn how to get started with the Hume API

Get started with the API

Discuss topics and questions with our developer community

Join community

An interactive podcast experience

Chatter

Your personal AI with emotional intelligence

Hume iOS App

Talk to an empathic AI

Hume App

Try our demos

Start building with EVI 2, our latest foundational voice-to-voice model with EQ

Empathic Voice Interface

Vocal burst

Facial expression

Language

Speech prosody

Sign up to generate over 100+ data points using our expression models

Get insights on your files

Please adjust your browser settings to allow Hume.ai to use your camera and microphone.

Your browser has blocked access to your camera and microphone.

Your camera and microphone are required to use this feature. Grant permissions to continue.

Allow Hume.ai to use your camera and microphone.

Hume models distinguish between 48 and 53 distinct emotion categories. Here we've selected a small set of basic emotions. Please note that these expression levels are not personalized. The minimum and maximum intensities used here are based on a large representative sample of expression data we have collected. The magnitude displayed in these levels may not be perfectly calibrated to your personal perceived level of each expression.

Expression levels

It looks like you've stopped interacting with the playground. The stream will automatically shut down after 5 minutes of inactivity.

Would you like to continue?

Hume models distinguish between 48 and 53 distinct emotion categories.  This visualization shows top expression detected in your speech prosody over time.

Speech prosody activity

Hume models distinguish between 48 and 53 distinct emotion categories. Here we show just the three categories with the highest intensity.

Top expressions

Hume models distinguish between 48 and 53 distinct emotion categories. This visualization shows the top expression detected in your vocal bursts over time.

Vocal burst activity

Text file analysis in the playground is coming soon!

Something went wrong while trying to download the predictions. Please try again.

Unable to download predictions

Inference

Build and customize models to suit your needs

Training

For businesses with high volume and advanced data control requirements

Enterprise

Ideal for individual developers, startups, and businesses that prefer a flexible pricing structure based on usage

Pay as you go

Pricing

Files are processed with Hume expression model APIs. Do you want to continue?

Please remove the following files from your uploads.

Validation errors

We were unable to load the chat history. Please try again later.

An unexpected error occurred

Hume will not save chat history for any conversations that are held while data retention is off. Please turn on data retention on <accountPageLink>your account page</accountPageLink> to preserve your chat history.

Data retention is currently off for this account

We were unable to load your chat history. Please try again later.

We were unable to load your voice configurations. Please try again later.

We were unable to load your tools. Please try again later.

We were unable to load your custom voices. Please try again later.

Chats

Customize and deploy a voice client using our Empathic Voice Interface that meets your specific needs

Create a custom voice client

Examples

A brief description of what the tool does

The content that will be used if the tool fails

{numConfigs, plural, =1 {1 invalid configuration was} other {# invalid configurations were}} found in your account which are not included in this list.

Invalid configurations found

{numVoices, plural, =1 {1 invalid voice was} other {# invalid voices were}} found in your account which are not included in this list.

Please adjust your browser settings to allow Hume.ai to use your microphone

Your browser has blocked access to your microphone

Summary will be available when connection is closed

There were not enough messages in your latest call to provide summary metrics on. Try starting a new conversation to see metrics.

Call summary

Untitled

We were unable to load preset voice list. Please try again later.

We were unable to load your custom voice data. Please try again later.

We were unable to load voice parameters. Please try again later.

Log in