AI powered logs

General-Purpose Log Viewer with AI Support

Done in one 3 days: A universal log viewer that brings all your log files into one place — regardless of format. Analyse large log files with AI, navigate entries with rich details, and extend the system with new log types without touching any main code.

Log Viewer

Key Features

  • AI Analysis: Send any log to an AI assistant for automated analysis, anomaly detection, and AI-generated filter expressions
  • Pluggable Type System: Add support for new log formats by dropping in a parser and renderer — no changes to the main application
  • Large File Support: Handles logs with more than 100,000 entries via background ingestion into SQLite or MySQL
  • Multi-Source Merging: Combine multiple files or entire folder trees into a single unified log view
  • Rich Filter Syntax: Expression-based filtering with field comparisons, contains, in [...], regex, and boolean logic; filters are saved per log
  • Detail Panel: Type-specific rich layouts with tables, collapsible pipeline steps, and highlighted error lines
  • Multi-Select & Copy: Range-select entries with keyboard shortcuts, copy to clipboard, and context menu
  • Themes: Light and dark UI themes

Technical Architecture

  • Backend: PHP 8+ with a pluggable type → parser → renderer pipeline
  • Storage: SQLite (zero-setup) or MySQL; background ingest worker keeps the DB continuously up to date
  • Type Detection: Automatic log type detection with optional URL parameter override
  • Frontend: Vanilla JavaScript with Fetch API — no heavy frameworks
  • AI Integration: Configurable LLM backend for log analysis and filter generation
  • Multi-Tenant Ready: Per-user and per-organisation settings, saved filters, and custom type overrides

Tech Stack

PHP • SQLite • MySQL • YAML • Vanilla JavaScript • Fetch API • AI / LLM

Built-in Log Types

  • TradingLog: Pipe-delimited order records with BUY/SELL badges, status pills, and latency highlighting
  • BatchJobLog: One file per job run with header block and timestamped lines; colour-coded pass/fail border
  • OrderPipeline: One subfolder per order with summary JSON and numbered step files; collapsible steps with auto-expand on failure
  • PhpErrorLog: Standard PHP error log with grouped stack traces and colour-coded severity pills

Use Cases

Engineers and DevOps teams can point the viewer at any log file or folder — trading systems, batch jobs, order pipelines, PHP errors — and immediately filter, navigate, and analyse entries. The AI assistant handles the heavy lifting: "Show me all rejected orders with a latency above 100 ms" generates the filter expression automatically.