Live Feed/PostHog/Fact Record
PostHog logo
PostHog
feature 96% Confidence Gate September 9, 2026

Multimodal models need video. We froze time to give it to them

PostHog introduced a new capability to convert session recordings into structured, multimodal-ready data formats. This allows LLMs to ingest and analyze user session video data directly for automated insights.

Verified State Diff

Comparison Mode:
- Previous State
Session recordings were stored as raw, unstructured video files that required manual human review to extract insights.
+ Verified New State
Session recordings are now structured and accessible for ingestion by multimodal AI models for automated analysis and pattern recognition.

Impact & Verification Analysis

WHO IS AFFECTED

Product managers, data analysts, and developers using PostHog for user behavior analytics.

WHY IT MATTERS

It significantly reduces the time-to-insight for user experience research by automating the analysis of session recordings, transforming a high-volume storage cost into a high-value data asset.

Full Fact Overview

PostHog is addressing the 'dark data' problem of session recordings by enabling the transformation of raw video playback data into a format compatible with multimodal large language models. By 'freezing time' and structuring these recordings, the platform allows developers to programmatically query user behavior patterns that were previously only accessible through manual observation. This integration bridges the gap between unstructured visual session data and AI-driven analytics, effectively turning dormant storage into actionable intelligence.

Multi-Source Evidence Chain (1)

Multimodal models need video. We froze time to give it to themPostHog
TRACKED ENTITY
Explore all historical PostHog changes
View PostHog Hub ➔