Detect what transcripts can't capture

Send video or audio, get back structured JSON: social signals with explanations, an engagement level, and conversation quality scores on request.

Get API key Read docs

One call, three layers of intelligence

Social signals and an engagement level come back by default. Add conversation quality scores when you want them - on the upload endpoint and the stream alike.

Layer 1Social Signals

What notable moments happened?

Agrmt

Intrst

Hesit.

Confid

Layer 2Engagement Level

How attentive was the person?

Engagement status

Layer 3Conversation Quality

How good was this interaction?

CQI

Clarity

Energy

Rapport

Layer 1Social Signals

What notable moments happened?

Agrmt

Intrst

Hesit.

Confid

Layer 2Engagement Level

How attentive was the person?

Engagement status

Layer 3Conversation Quality

How good was this interaction?

CQI

Clarity

Energy

Rapport

Every signal your AI needs to read the room

Each signal includes a probability level and a human-readable rationale explaining what triggered it. Inter-2 reports all ten; Inter-2 Audio reports six of them from speech alone: confidence, interest, agreement, hesitation, uncertainty and frustration.

Agreement

Alignment with another person's position, intent, or understanding

Confidence

How firmly and assuredly someone communicates their position or decision

Confusion

A breakdown or gap in understanding during an interaction

Disagreement

Active divergence from another person's viewpoint or proposal

Frustration

Mounting tension or irritation when progress toward a goal feels blocked

Hesitation

Uncertainty or delay before committing to a response or action

Interest

Attention or curiosity toward something unexpected or stimulating

Skepticism

Questioning or doubtful stance toward a claim, proposal, or explanation

Stress

Heightened tension or unease during an interaction

Uncertainty

A temporary disruption in speaking or responding to a question

Agreement

What it detects

Alignment with another person's position, intent, or understanding

API response

{
  "signal_type": "agreement",
  "start": 1.2,
  "end": 8.7,
  "probability": "high",
  "rationale": "The speaker explicitly states, "Yeah, that was great," providing clear verbal confirmation of his stance. He also nods his head while speaking and maintains steady eye contact with the camera, demonstrating active participation and alignment with the topic."
}

Use cases

Detecting buy-in during sales calls, confirming understanding in training, measuring alignment in negotiations

Engagement, always reported

Unlike signals such as hesitation or agreement, engagement is always reported: engaged, neutral, or disengaged, sent the moment it shifts.

Engagement status

0:000:52

Engaged

Neutral

Disengaged

Stream events

{ "type": "engagement.updated", "data": { "state": "engaged", "start": 0 } }
{ "type": "engagement.updated", "data": { "state": "neutral", "start": 12.3 } }
{ "type": "engagement.updated", "data": { "state": "engaged", "start": 18.5 } }
{ "type": "engagement.updated", "data": { "state": "disengaged", "start": 35.2 } }
{ "type": "engagement.updated", "data": { "state": "engaged", "start": 38.8 } }

Interaction quality, scored automatically

The API computes a Conversation Quality Index (CQI) from the detected signals - a single 0-100 score, broken down into 5 behavioral dimensions. Request the whole-session score, a timeline that updates through the conversation, or both.

CQI

Clarity

Authority

Energy

Rapport

Learning

Clarity

How easy the speaker is to follow, reflecting the organization, concision, and coherence of their ideas.

Authority

How confident, decisive, and credible the speaker comes across in their delivery.

Energy

The speaker's level of vitality and active engagement throughout the interaction.

Rapport

The warmth and emotional quality of the interaction, reflecting how acknowledged and at ease the other person feels.

Learning

How openly the speaker reflects, tests ideas, and adapts, capturing their curiosity and growth orientation.

Powered by Inter-2

Our social intelligence model, built to read human social signals as an interaction unfolds.

Real time

Inter-2 reports each signal as it starts and ends, so a product can respond while the conversation is still happening. Upload a recording when you need a batch read instead.

Video or audio

Inter-2 reads what it sees and hears together. Inter-2 Audio reads confidence, hesitation and uncertainty from speech alone when there is no camera.

Explainable

Every signal carries a rationale naming the cues behind it, so you can audit a detection or pass it to an LLM as context.

Not sentiment analysis

Sentiment gives you positive/negative. We give you 10 specific signals with timestamps and rationale.

Not DIY multimodal

Don't rely on a fragile patchwork of disparate models. Inter-2 processes the modalities together and delivers insights through one API.

Not a black box

Every signal includes the observable cues that triggered it - which modalities, which behaviors. Auditable and actionable.

What you get from every API call

Social signals, engagement and quality scores, with the rationale behind each, over REST or WebSocket.

10 social signals

Agreement, confusion, frustration, hesitation, and 6 more with Inter-2 - each with a probability level and a rationale explaining the observable cues that triggered it. Inter-2 Audio reports six of them from speech alone.

Multimodal analysis

Video and audio processed together in temporal alignment. Gaze direction, vocal prosody, posture, gestures, and speech - not just transcripts.

Engagement tracking

Engaged, neutral, or disengaged, sent the moment it shifts - so you always know how oriented a person is to the interaction.

Conversation Quality Index

A 0-100 score derived from social signals across five behavioral dimensions - Clarity, Authority, Energy, Rapport, and Learning. Request it for the whole session, over time, or both.

Explainable by default

Every signal includes its rationale, with the modalities and behaviors involved. Auditable, actionable, and ready to forward to your LLM.

REST and WebSocket

Upload a recording to the REST endpoint, or stream live video and audio over a WebSocket and receive signals as they start and end. Authenticate with your API key or a short-lived token.