Stop ignoring your session replays.

VES AI watches every PostHog session replay and turns them into rich user stories, product insights, and qualitative learnings your team can actually act on.

$

The replay graveyard

Your PostHog replays are a gold mine. Nobody's watching them.

<1% watched

Your team records thousands of sessions. Almost none get reviewed.

Hidden bugs

Real issues hide in replays nobody has time to watch.

Vibes-based decisions

Product decisions get made on gut feel, not evidence from real user behavior.

AI that watches every session

From individual user stories to account-level intelligence to product research — all from your existing PostHog data.

How is bryce@company.com doing?

$ vesai user bryce@company.com
{
  "email": "bryce@company.com",
  "sessionCount": 8,
  "averageSessionScore": 71,
  "userScore": 65,
  "health": "At risk — recurring friction in core workflow",
  "story": "Power user exploring advanced features across 8 sessions over 7 days. Engagement is growing — session duration increased 40% week-over-week and feature breadth expanded from 3 to 7 unique features used. However, a critical blocker emerged: 3 of 8 sessions ended at the same CSV export timeout when datasets exceeded 10k rows. The user attempted workarounds (filtering down, paginating) but ultimately abandoned each time. Separately, filter state resets on back-navigation forced repeated configuration in every session, adding ~2 minutes of friction per visit. This user is deeply invested in the product but hitting real walls that will drive churn if unresolved.",
  "sessions": [
    {
      "sessionId": "s_a1b2c3",
      "score": 82,
      "activeSeconds": 342,
      "features": ["dashboard", "export", "filters"],
      "issues": ["Export timeout on large dataset"]
    },
    {
      "sessionId": "s_d4e5f6",
      "score": 45,
      "activeSeconds": 89,
      "features": ["export"],
      "issues": ["Export timeout", "Rage-clicked export button 4x"]
    },
    ...
  ],
  "markdownPath": ".vesai/workspace/users/bryce-company-com.md"
}

How it works

PostHog

Connect

Pull session replays and events from your PostHog project

Playwright

Render

Replay each session in headless Chromium via Playwright

Google Cloud

Analyze

Gemini watches each video and extracts structured insights via Vertex AI

Act

User, group, and research intelligence ready for your team or your agents

Built for your AI agents.
Works for humans too.

JSON-first output

Every data command emits structured JSON. Pipe output directly into your agent workflows, CI pipelines, or dashboards.

Ships with a SKILL.md

A comprehensive skill file teaches Claude Code, Codex, OpenClaw, and other coding agents how to use every vesai command out of the box.

Durable markdown artifacts

User stories, group analyses, and research answers persist as markdown in your project’s .vesai/workspace/ — greppable, portable, version-controlled.

Local-first, no SaaS

Your PostHog keys, your GCP project, your machine. No data leaves your infrastructure. No vendor lock-in.

Getting started

  1. 1Install VES AI
    curl -fsSL https://ves.ai/install | bash
  2. 2Global setup
    vesai quickstart
  3. 3Initialize your project
    vesai init
  4. 4Analyze your first user
    vesai user bryce@company.com