Runs in your browser · Nothing is uploaded to a server

When words deceive, the face still speaks

Firasa reads emotions, facial movement, expression and the voice. You can see how open a person is, because the body, the eyes and the voice say more than words do — down to micro-expressions that last a fraction of a second and strain in the voice the ear does not catch.

Free, no card required. The camera turns on only when you press the button.

Pricing

What it can do

All of it runs on your own device — with a network connection and without one.

Real-time analysis

The camera starts when you press the button, and the analysis runs as you record. On screen: the emotion, how strongly it shows, and the list of signs — which facial muscles fired. Beside it, a chart of the last thirty seconds shows how the state shifted. The first few seconds are spent learning your ordinary face; from then on the app measures deviation from that, not from an average table. One button captures a photo and analyses a single frame.

Micro-expressions

A flash lasting 40 to 500 milliseconds is a tenth or two of a second — too fast for the eye. The app watches each muscle group separately and catches the moment the movement peaks: it saves the frame at exactly that instant and records the time. The report shows how many flashes there were, when each happened, and which emotion appeared.

Privacy comes first

There is no server that could receive your video — we did not build one. The recognition model downloads into the browser once and runs on your device from then on. The history of analyses lives in the browser itself: any entry can be deleted and the whole history cleared. The database holds only the account: email, name and avatar.

Detailed reports

A PDF with a colour key, charts of emotions and states, and frames from the key moments with timestamps. At the end, a conclusion in plain words: which emotion dominated, whether the smile involved the eyes, how many flashes there were. The report can be attached to your work or shown to a client.

Training

Seventy training lessons: an expression flashes for a fraction of a second and you name the emotion. Difficulty rises, progress is saved, mistakes are explained. Alongside them, an explanation of the method itself and an assistant that answers questions about how it all works.

What makes Firasa different

  • Analyses in real time — you see emotions while you are still recording

  • Catches micro-expressions — flashes of 40–500 milliseconds that the eye cannot catch

  • Hears the voice — pitch, tremor, pauses and pace — strain is audible before it shows on the face

  • Weighs how open the expression is — does the smile use only the mouth, or the eyes as well?

  • Privacy comes first — the video stays on your device and is sent nowhere

  • Works offline — the internet is needed only while the model downloads

Who it is for

Firasa is used by:

Illustrative scenarios, not real analysis results. Record with participants’ consent. Facial movement does not establish honesty, danger, diagnosis, fitness to fly or credit eligibility.

What Firasa does not do

This is a learning tool, built within the law. It does not detect lies, does not diagnose and passes no judgement on anyone. The app shows observable signs and how strongly they appear; the conclusion is always yours to draw.

Try it right now — free

Turn on the camera and see what your own face shows. Signing up takes half a minute.