v1.1 Specification Published

eitp://

A protocol for transferring behavioral-emotional context from web telemetry to LLMs. Eliminate disambiguation. Remove the awkward from every chat window.

Read the Specification Try the Playground
The Problem

LLMs start every conversation blind

Users browse for minutes before engaging. They scroll past pricing, hesitate on CTAs, switch tabs to compare. All of this signal is discarded when the chat window opens. The LLM guesses from scratch.

The Mechanism

Behavioral telemetry encodes emotion

Mouse kinematics (velocity, acceleration, jerk), scroll patterns, click dynamics, proximity to interactive elements, and tab behavior carry emotional signal. EITP captures it, structures it, and transmits it.

The Result

Context replaces guessing

The LLM receives a nine-dimensional packet — emotion, spatial, temporal, behavioral, patterns, intent, page content, global intelligence, and channel origin — and responds with relevance from the first interaction.

Production Evidence

From the reference implementation across automotive, RV, and powersports verticals

62%
Convert under distress
Uncertainty or cognitive overload
2.7x
Harder path
Distressed vs. engaged buyers
22
Emotion taxonomy
Production-validated states
9
Packet dimensions
Structured context per session

Protocol Layers

Signal Capture 8 tracker modules capture raw behavioral events | Transport Events batched and transmitted via WebSocket | Filtering Bot detection removes non-human sessions | Inference Telemetry transformed into emotional classification | Packet EITP v1.1 packet assembled (8 dimensions) | Consumption Packet delivered to LLM as context
Read the full specification

See It Live

SeerChat is the reference implementation of EITP — a production system reading emotional context across automotive, RV, and powersports verticals.

Visit SeerChat