v2.0 — Stream Intelligence, Full Automation & PySide6 Rewrite

⚡ Kick Stream Analytics & Deep AI Research Tool v2.0

A complete rewrite from tkinter to PySide6 (Qt6), with a major set of new
detection and automation features.

🆕 What's New

• 📊 Stream Intelligence — historical fraud analysis across days/weeks
  of monitoring. Fraud score timeline, full-year risk calendar, pattern
  detection (escalating, persistent, scheduled botting, etc), cross-streamer
  leaderboard, and "Report to Kick" evidence generator.
• 🚀 Full Automation — fully unattended monitoring. When a tracked
  streamer goes live, the app opens a session, starts chat logging, runs AI
  analysis every 30 minutes, and auto-saves reports — no interaction needed.
• 🔊 Stream Audio Monitor — listen to any live stream's audio while
  monitoring, no browser required.
• 🤖 AI Analysis Queue — multiple sessions can queue for the same AI
  provider with automatic cooldowns. Every report auto-saves to AI_Reports.
• 🔧 API Diagnostics — 10-test connection health check for Kick,
  Anthropic, and Ollama from the hamburger menu.
• 🎯 Improved detection accuracy — spike/chat correlation signal,
  accurate gift sub counting, subscription deduplication, raid context
  passed to the AI, and a CONFIRMED BOTTING verdict tier.

🚀 Getting Started

1. Download the source .zip below (or clone the repo)
2. Run install.bat — sets up Python, packages, and ffmpeg automatically
3. Run launch.bat to start the app

Want a standalone EXE instead? Run build.bat after install.bat —
output goes to dist\KickAnalytics_v2\, fully self-contained, no Python
needed to run it.

📖 Documentation

• README.md — full feature overview
• AI Guide.txt — AI model setup and reading analysis reports (also built
  into the app)
• VIEWBOT_EXPLAINED.txt — plain-English explainer on how viewbot/chat bot
  fraud works and how this tool detects it

⚠️ Status

This release is in active testing across multiple machines. Please report
any issues via GitHub Issues.

---

This tool is not affiliated with or endorsed by Kick.com

Kick Stream Analytics v1.5

What's Changed in v1.5

New Features

• 👁 Hidden viewer count detection — detects when a streamer
  has enabled Kick's "Hide Viewer Count" feature
• Side panel shows last known count with 👁 Hidden indicator
• AI prompt automatically warns the model when viewer count
  is hidden and adjusts analysis accordingly

Improvements

• Channel Overview polling increased from 30s to 10s
  for better viewer spike detection and AI curve analysis
• AI max tokens increased from 3000 to 3500 for both
  Ollama (local) and Anthropic Claude API
  — longer more complete analysis reports

Bug Fixes

• Fixed viewer count displaying as — (null) for some streamers
• Fixed hidden viewer count not being detected correctly

Kick Stream Analytics v1.4

What's Changed in v1.4

New Features

• 💾 Chat Logs tab — record live chat sessions to .txt files
• ▶ Start / ■ Stop recording in Live Chat tab toolbar
• ☑ Auto Log — auto-starts recording per streamer (state saved per streamer)
• 🔴 REC pulse indicator in session topbar when recording
• Auto Log column in Multi-Channel — synced with Live Chat Auto Log
• 3 username filter boxes in Chat Logs tab
• Keyword search box — searches message content
• @ Mentions filter box — filters messages containing @
• 💬 Replies filter box in both Live Chat and Chat Logs tabs
• Chat Log panel shows full unfiltered log
• Filtered Chat panel shows all filtered results
• Save folder persists across app restarts
• Auto-creates streamer subfolder for organised saves
• System folder protection when selecting save location

Bug Fixes

• Fixed AI verdict incorrectly showing Suspicious when report says Clean
• Fixed verdict parser matching keywords in explanatory sentences
• Fixed manual Start recording creating empty 0kb files
• Fixed Load Chat Log not opening in streamer subfolder

Changes

• Chat Logs tab moved between Report and AI Guide tabs
• Auto Log renamed from Auto Save
• Manual Save button removed
• Start/Stop/Auto Log moved to Live Chat toolbar

Kick Stream Analytics v1.3

What's Changed in v1.3

New Features

• 🔔 Live Notifications — popup alerts when tracked channels go live
• 🤖 Auto Mon — automatically start monitoring when a channel goes live
• 👥 Followers Only mode added to side panel Chatroom section
• 🔗 Social links in Streamer Info are now clickable
• 🔄 Auto Scan added to Chatters (30s), Top Words (20s), Activity Chart (20s)
• 🎙 Streamer Info auto-loads when monitoring starts

Bug Fixes

• Fixed all session start issues from Top 100, Multi-Channel and notifications
• Fixed followers_mode API crash on session start
• Fixed social links not opening browser
• Fixed ChatMoveToSupportedChannelEvent showing as UNKNOWN
• Fixed pusher:error and ChatroomUpdatedEvent showing as UNKNOWN

New Events with Icons

• ⭐ SUB — ChannelSubscriptionEvent now correctly captured
• 🎁 GIFT — LuckyUsersWhoGotGiftSubscriptionsEvent now correctly captured
• ✅ UNBAN — UserUnbannedEvent
• 📴 STREAM ENDED — StopStreamBroadcast
• 🗑 DELETED — MessageDeletedEvent

License Update

• Changed from MIT to Personal Use Only license

Kick Stream Analytics v1.2

What's Changed in v1.2

Bug Fixes

• Fixed missing event types causing undercounted subs and gifts in AI analysis
• ChannelSubscriptionEvent now correctly captured and counted
• LuckyUsersWhoGotGiftSubscriptionsEvent now correctly captured

New Events with Icons

• ⭐ SUB — ChannelSubscriptionEvent
• 🎁 GIFT — LuckyUsersWhoGotGiftSubscriptionsEvent
• ✅ UNBAN — UserUnbannedEvent
• 📴 STREAM ENDED — StopStreamBroadcast
• 🗑 DELETED — MessageDeletedEvent

New Features

• Auto Scan added to Chatters tab (30s, ticked by default)
• Auto Scan added to Top Words tab (20s, ticked by default)
• Auto Scan added to Activity Chart tab (20s, ticked by default)
• Streamer Info tab now auto-loads when monitoring starts
• Streamer Info Load Info button removed — fully automatic

Kick Stream Analytics v1.0a

Kick Stream Analytics & Deep AI Research Tool — v1.0

Initial public release.

What's Included

• Live Kick.com chat monitoring via Pusher WebSocket
• Multi-stream support — monitor multiple streamers simultaneously
• AI-powered viewbot detection (Ollama local + Anthropic Claude API)
• GPU auto-detection with model recommendations
• Real-time viewbot scoring panel
• Top 100 global streamer rankings
• Multi-channel monitoring with auto-refresh
• Chatter classification and spam detection
• Full session reports with export
• Streamer profiles, VOD history and clip browser
• Built-in AI Guide tab

Two Ways to Run

• Source version — run install.bat then run.bat (zero AV flags)
• Build your own EXE — run build_folder.bat to compile

Requirements

• Windows 10 or 11 (64-bit)
• Python 3.9 or higher
• Ollama (optional — for free local AI analysis)

Notes

• No Kick account or API key required for core features
• See README for full setup instructions and AV notice

Kick Stream Analytics v1.01a

What's Changed in v1.01a

Documentation

• Added README.md with full GitHub formatted documentation
• Added screenshots folder with app preview images
• README.md includes feature list, setup instructions, GPU tier table,
  tab overview, package list and antivirus notice