Pre-trained Emotion Recognition Across Face, Voice, and Text
Production-ready emotion models for face, voice, and text — benchmarked, calibrated, and ready to drop into your stack. ONNX, TFLite, and server-side variants.
3 modalities
face · voice · text
ONNX + TFLite
edge + server
Benchmarked
open evaluation set
Proof Block
What ships is measurable and inspectable
Not just an abstract emotion score. The model output is structured enough to audit, calibrate, and wire into downstream policy.
Live FER demo
Live browser demo available
Try real-time facial-expression capture in the Voice Agent demo on this site.
Output schema
emotion + confidence + valence/arousal + AUs
Single forward pass returns both category and dimensional affect.
Deployment shape
edge + server variants
The same label space carries across ONNX, TFLite, and hosted inference.
Example response shape
Sample model output
{
"emotion": "joy",
"confidence": 0.93,
"valence": 0.71,
"arousal": 0.38,
"action_units": {
"AU6": 0.82,
"AU12": 0.91
}
}JSON
Three modalities. One coherent label space.
Designed to be fused — see also Multimodal Fusion for the joint model.
Where it fits
🎮
Games
NPC reactions tied to real player emotion.
🤖
Assistants
Emotion-aware response generation.
📞
Contact centres
Real-time sentiment for live calls.
📱
Consumer apps
On-device emotion read for mobile features.
Build with emotion that actually generalises
Request developer access — model cards, benchmarks, and integration guides included.
Get developer access →Talk to David