No glitches or transcription errors detected. All systems nominal! 🟢
No pending link updates. All links in sync! 🟢
📜System Activity Terminal Logs
Web Server: ACTIVE
AI Core: CHECKING...
⚡ CPU LOAD0%
💾 MEM USAGE0%
Isolate log module:
Category Filter:
Timestamp
Category
User (Role)
Client IP
Action
Details
Loading persistent audit logs database...
🧹 Guest Session Cleaner
Perform administrative cleaning sweeps on the local sqlite and cloud database records to wipe out stale guest users.
Directory Filter:
User / Guest Identity
Role
IP Address History
Last Active
Total Pings
Plays
Linked Identities (Heuristics)
Loading user profiles directory...
Audio Streams & Playback Metrics Dashboard
Total Shows Tracked
0
Unique Listeners (UUIDs)
0
Total Play Sessions
0
Episode Show Title
Catalog Tier
Unique Listeners
Total Plays
Peak Play Offset
Loading playback statistics...
🏆 Curated AI Rankings / Lists
Regenerate the global funniest list and ranks across the site using the AI models' compiled details.
Processing...0%
🔄 Database Catalog Scraper
Re-scan and update the active shows catalog from remote database endpoints.
📤 Import Folder Scanner
Scan for raw local transcript files on the server to import.
💾 Export / Import Database JSON
Back up or restore full database JSON profiles directly.
🧪 AI Analysis Prompt Sandbox
Tweak, test, and save the AI analysis prompt template system-wide.
Prompt Preset:★ ACTIVE LIVE
📡 Existing Live Analysis
No show selected.
🧪 Sandbox Test Output
Click "Run Test" to run the edited prompt template.
Browse historical sandbox test runs, inspect customized inputs/outputs, and restore past versions.
Select a test run from the sidebar to inspect its prompt template, parameters, and AI model output.
⚙️ System-Wide AI Engine Settings
Configure the pipeline provider models, transcription chunking limits, and default overlays. These settings apply to all newly initiated analyses.
🎛️ Dual-Engine Architecture (Manual & Auto Bot)
⚡ Zero Silent Fallback Mode Active
🧠 Manual Runs (Single Show)Interactive
🤖 Automatic Pipeline (24/7 Bot)Daemon
Default: 1000 segments. Smaller sizes consume more AI calls but capture denser shows better.
Default: 25%. Ensures calls occurring on chunk boundaries are not lost.
⚡ Transcript Optimization & Token Reduction
✂️ 5m Pre-Show Trim
Excludes 00:00–05:00 intro banter & music where no calls occur, saving ~3,000 tokens per show.
⏱️ 15s Subtitle Grouping
Compresses rapid subtitle lines into 15s windows, delivering -41% token reduction with precise sync.
⛓️ Single-Pass 1M Context
1,048,576 token capacity ingests entire 3-4h shows at once, eliminating chunk boundary loss entirely.
🛡️ Zero-Loss Safeguard
Strict API routing with checkpoint_analysis.json; zero silent fallbacks on error.
Controls active steps in the AI pipeline by dynamically enabling/disabling the corresponding templates on disk.
When enabled, importing a Deepgram Cloud transcript will automatically replace any existing local Whisper transcript to eliminate hallucinated text.
Only newly discovered shows posted on or after this date will be automatically transcribed via Deepgram Cloud. Default: 2026-06-04.
Warn on Discord when balance drops below this. Default: $10.00.
Auto-pause the bot when balance drops below this. Default: $5.00.
◫ Inspector Default View Layout
Default new inspector overlays to Chunks segmented tabs layout when chunk files exist.
🤖 Auto-Transcribe New Episodes (Deepgram)
Automatically run Deepgram transcription when new shows are discovered.
🧠 Auto-Analyze New Episodes (AI Checker)
Automatically run AI checklist/analysis after a show is transcribed.
🔑 API Key Health Tracker
Rotated Gemini API keys status. Dead keys are automatically skipped and pinged to Discord.
💰 Cost & Token Audit Logs
Real-time token metrics and approximate API costs in USD per show/episode run.
Episode
Provider
Tokens (In/Out)
Time
Cost (USD)
Status
Log File:
Select a log file above to preview...
💾 Storage & Cache Metrics
VPS disk space status and transcript/audio segment caches size.
VPS Disk Space
0.0 / 0.0 GB
0% Used
Cache & Content Size
0.0 MB
Transcripts, segments & databases
⚠️ **Warning**: VPS free storage is below 10%! Please purge database backups or logs.
AI Analysis Inspector
Show ID: default
View Run:
📥
Drop Transcript JSON Here
Must be a valid transcript JSON file with metadata and segments
📥 Bulk Export Runs
Select the runs of this show you wish to download as a ZIP file containing prompts and responses.
0 runs selected
⚡ Mission Control: Live Operations & Telemetry
Idle
Currently Processing Episode
No active show running
System is standby.
Elapsed Latency
⏱️ 0.0s
Overall Progress
0%
9-Step Pipeline Lifecycle
Real-Time LLM Token & Cost Metrics
Input Tokens0
Output Tokens0
Reasoning Tokens0
Estimated USD Cost$0.0000
# No active API traces available for current operation
# No LLM response chunks recorded yet
📟 Worker Activity Logs
0 lines
Waiting for live operations...
📋 Upcoming Operations Queue (0 shows queued)
Queue is currently empty.
🧠 AI Bulk Analysis Manager
Bulk scan the database and batch queue shows for LLM prank call analysis.
0
🟢 Analyzed
0
🟠 Pending AI Check
0
📂 No Transcript
📊 Preset Distribution:
0 candidates matching
0 shows selected for queue
Select
Show Title
Year
Tier
Prompt Preset
AI Status
🎙️ Deepgram Cloud Transcription Bulk Manager
Bulk scan the database and batch queue shows for Deepgram Nova-2 cloud transcription.
0
☁️ Cloud Deepgram
0
💾 Local Whisper
0
❌ No Transcript
0 candidates matching
0 shows selected for queue
Select
Show Title
Year
Tier
Current Engine
⚠️
Deepgram Cloud Operations Info:
This bulk transcription uses Deepgram Cloud API. The daily 24h limit checks are bypassed since this is a manually approved administrative request. Shows will download, transcribe, and replace their local transcripts in background thread sequentially.
🧹 Database Reset & Export Manager
Compile ZIP exports of transcripts + prank calls, or clean up bad local/Whisper transcripts & calls.
0
☁️ Cloud Deepgram (Protected)
0
💾 Local Whisper (Clearable)
0
❌ No Transcript
📊 Preset Distribution:
💡Your progress is safe! Custom episode notes, completions, hearts/likes, and call notes are stored securely in settings, so they are fully preserved during resets. The player automatically re-aligns them using dynamic time-tolerance (up to 15s) even if audio timestamps shift slightly after re-transcribing.
0 candidates matching
0 shows selected
Select
Show Title
Year
Tier
Prompt Preset
Current Status
🚨 LOCAL RESET ONLY:
⚙️ System Settings Manager
Modify environment variables, Discord webhooks, and AI analysis models dynamically.
🔍
🔑 API Keys Rotation Pool
Gemini Pool Size:
Add New API Key
Google Gemini Free Tier Mode
Rotate ONLY Free Tier keys in the pool. When OFF, rotates Paid Tier keys.
🔐 Master Admin Credentials & Server Settings
🎛️ Provider Controllers
💡 For dedicated runtime parameters, model spec cards, and 24/7 automation rules, see the 🧠 Manual AI Runs and 🤖 Auto Bot Runs tabs.⚡ Zero Silent Fallback Active
🧠 Manual AI Runs & Model Selection
⚡ Zero Silent Fallback Mode Active
Default when clicking 'Run AI Analysis' on any episode. Absorbs full show in a single pass.
Cross-chunk reconciliation and deep multi-pass synthesis engine.
✂️ 5m Pre-Show Trim
Excludes 00:00–05:00 intro banter & music where no calls occur, saving ~3,000 tokens per show.
⏱️ 15s Subtitle Grouping
Compresses rapid subtitle lines into 15s windows, delivering -41% token reduction with precise sync.
⛓️ Single-Pass 1M Context
1,048,576 token capacity ingests entire 3-4h shows at once, eliminating chunk boundary loss entirely.
🛡️ Zero-Loss Safeguard
Strict API routing with checkpoint_analysis.json; zero silent fallbacks on error.
⚙️ Pipeline Execution Settings
0.0 = deterministic extraction; 0.7 = creative.
Default: 24,000 characters.
Default: 15% overlap between chunk boundaries.
Enable Chunk Mode
Split long episodes into parts
Enable Multi-Run Strategy
Process multiple passes for deep synthesis
🤖 24/7 Automation Daemon Pipeline Settings
🛡️ Daemon Watcher Active
Runs automatically on new shows discovered by Deepgram.
Autonomous merge & synthesis pass for queued background runs.
🔄 Auto Discovery & Queuing
Polls feeds every 60s. Enqueues newly published episodes on or after the cutoff date.
⏱️ 24h Safe Quota Limit
Caps unattended runs per 24h period (default: 2/day) to guarantee safe budget enforcement.
✂️ Auto 5m Trim & 15s Grouping
Employs -41% token optimization automatically before feeding into the analysis model.
🚨 Alerts & Checkpoints
Preserves intermediate state in checkpoint_analysis.json; sends Discord webhook on warnings.
Enable Deepgram Auto-Transcription
Automatically transcribe new episodes using Deepgram
Enable AI Checker Auto-Analysis
Automatically run AI analysis on new transcripts
⚙️ Auto Bot Pipeline Execution Settings
Authoritative execution parameters applied autonomously by the 24/7 daemon watcher whenever a new show is transcribed.
0.0 = deterministic extraction; 0.7 = creative.
Default: 24,000 characters.
Default: 15% overlap between chunk boundaries.
Daemon Enable Chunk Mode
Split long episodes into parts for automated queue runs
Daemon Enable Multi-Run Strategy
Process multiple passes for deep synthesis on automated queue runs
Exclude 00:00–05:00 intro music & banter (~3,000 tokens saved).
Merge short fragments into 15s semantic blocks (-41% token reduction).
⚙️ Pipeline Pre-Processing & Chunking Strategy
Exclude pre-show music and banter where no prank calls occur (default: 5 min).
Merge short transcription fragments into structured blocks (default: 15s).
Single-pass leverages modern 1M+ token context windows for higher accuracy and lowest latency.
🛡️ Automatic Long-Show Chunking Guard
Automatically slices long shows exceeding duration or segment thresholds to prevent LLM cognitive attention collapse
Enforce multi-chunking for shows at or above this length (default: 6.0h).
Enforce multi-chunking when total transcript segments exceed this count (default: 5,000).
Segments per chunk window (~70–85 min of audio).
Overlap buffer percentage between consecutive chunks (default: 25%).
📊 Active Model Specification & Economics
🎙️ Deepgram Settings
⚠️ Unsaved Changes!
🚀 Pre-Flight Execution Briefing
Authoritative configuration & cost audit before launching analysis pipeline.