A product of createIF Labs GmbH, Tuningen
Humexa.
The digital assistant who knows your product range.
100 % local
Speech recognition, language model and voice run on the device in the store, and it still feels like a completely natural conversation with extremely low latency. No conversation goes to a cloud.
Only your knowledge
Humexa answers from what your company provides: product range, product data, company knowledge and, if you want, live data from your systems. What isn’t there, she says openly.
Natural interaction
She responds to her name, looks at you and answers with voice and lip movement. A “stop” interrupts her.
She answers in the language she is spoken to in. German, English and French are standard; on request we extend to 27 languages.
Simulated demo with sample content, audio in German
Privacy you can see.
Humexa works in public spaces, so privacy has to be visible to every visitor, not just written in a policy.
- The camera runs only during a conversation.
- On the display, visitors appear only as a silhouette.
- Nothing is saved until a spoken yes, then as text and for 90 days.
Only your knowledge.
Asked about something outside your store, Humexa doesn’t guess. That is the difference between an assistant and a chatbot from the internet.
How much is a flight to Rome?
I have no information on that. I only know what this store provides to me.
Speaks your customers’ language.
Deutsch English Français + 24
She answers in the language she is spoken to in. German, English and French are standard; on request we extend to 27 languages.
100 % customisable.
Look, voice, personality and knowledge, designed for your brand.
- Voice
- Your own voice through voice cloning, from a short recording.
- Person
- We scan real people in 3D, for example someone from your team.
- Scene
- We model the room freely to your wishes or scan your own in 3D.
- Character
- It doesn’t have to be a person: any character can become Humexa.
Optional, switched on only on request.
Humexa can perceive more if your use case calls for it. These features are off by default and are only activated at your request.
- Person recognition
- She recognises returning visitors, continues conversations and remembers what she knows about the person.
- Facial emotions
- She reads the mood from facial expressions.
- Vocal emotions
- She hears the mood in the sound of the voice.
- Motion recognition
- She recognises posture and gestures.
A whole conversation, from “Humexa” to “bye”.
The demo above as text. When a visitor switches language mid-conversation, Humexa switches with them.
- VisitorHumexa?
- HumexaHello! May I save our conversation so we can improve Humexa? Yes or no?
- VisitorNo.
- HumexaAll right, I won’t save anything. What are you looking for?
- VisitorWhere can I find basmati rice?
- HumexaYou’ll find basmati rice with the dry goods. I’ll show you the area on the map.
- VisitorDo you have oat milk?
- HumexaYes, in the chilled section. Let me show you on the map.
- VisitorHow much is a flight to Rome?
- HumexaI have no information on that. I only know what this store provides to me.
- VisitorThanks, bye!
- HumexaYou’re welcome. Goodbye!
Meet Humexa.
In a demo you experience Humexa live, and we discuss how she would work on your shop floor.
Developed and operated by createIF Labs in Tuningen, the same team that builds AI agents for production. To createIF Labs