Skip to main content
The Co-Therapy Agent is a silent multimodal partner — it never speaks to the patient. It runs the Vocal Acoustic Engine, the Facial Physiological Engine, and the Neurobehavioral Construct Computer to produce a canonical clinical payload with 15 named construct activations.

Pipeline

medera/cotherapy-quickstart.svgco-therapy quickstartCo-Therapy multimodal — quickstartOpen WSS /therapyStream audio + videoSubscribe to construct activationsHandle crisis eventswss://api.medera.info/therapy/Audio 16 kHz + Video 15 fps · multimodal fusionNeurobehavioral activations · facial physiology · vocal acousticsLive transcript · construct panel · crisis overlay
1

Create the session

2

Start the session

3

Analyze the session

Upload audio + video for full multimodal analysis. The analyzer runs the engines, computes the neurobehavioral profile, and (optionally) retrieves clinical RAG evidence.
4

Read the canonical payload

The response is a canonical payload with flat ClinicalMetric envelopes:
The Co-Therapy Agent never addresses or interacts with the patient. Recording requires explicit per-session consent (consent_recorded: true).

What’s next

Co-Therapy Agent

Agent-level documentation.

Multimodal architecture

Pipeline deep dive.

Neurobehavioral Construct Constructs

15 named constructs.

Multimodal API

REST surface.