Features LLM Export

LLM Export

Generate a structured markdown report of your connection status, optimized for analysis by AI tools like ChatGPT, Claude, or Gemini. Useful when you want a second opinion on whether your signal values indicate a real problem.

How to Use

1. Go to Data Export in the sidebar
2. Choose export mode:
   • Full (default): 48 hours of context, all data sources, incident journal
   • Update: Last 6 hours only, for follow-up questions
3. Click the export button
4. Copy the generated text and paste it into your preferred AI tool

What's Included

The export is a self-contained markdown document with tables, designed so an LLM can analyze your connection without needing additional context:

Section	Content
Overview	ISP name, tariff (down/up), current health status, active issues
Summary	Aggregated DS/US metrics: channel count, power (min/avg/max), SNR, error totals
Downstream Channels	Per-channel table: frequency, power, SNR, modulation, correctable/uncorrectable errors, DOCSIS version, health
Upstream Channels	Per-channel table: frequency, power, modulation, multiplex type, DOCSIS version, health
Events	Recent events (48h or 6h) with severity, type, and message
Speedtest Results	Last 10 results (full) or 3 (update): download, upload, ping, jitter, packet loss
Incident Journal	Full mode only: date, title, and description of documented issues
Cross-Source Correlation	Table linking speedtest results with signal health at the time of each test
Reference Values	DOCSIS thresholds from VFKD guidelines so the AI knows what "good" and "bad" means

Built-In Analysis Prompt

The export ends with specific questions for the AI:

1. Overall connection health assessment
2. Channels that need attention (with reasons)
3. Error rate analysis and whether it indicates a problem
4. Specific recommendations to improve connection quality

This turns your raw DOCSIS data into an actionable diagnosis.