Skip to Content

Source Reliability Export

Last modified by Robert Schaub on 2026/02/08 21:46

Source Reliability Export Guide

Version: 2.6.38
Last Updated: January 26, 2026

Overview

The Source Reliability admin page includes export functionality for source evaluation data. You can export individual source evaluations in multiple formats for documentation, review, or archival purposes.

Accessing Export Features

  1. Navigate to /admin/source-reliability
  2. Click the "View" button (eye icon) on any cached source entry
  3. The Source Evaluation Details modal opens with export buttons in the footer

Export Formats

Print / PDF

  • Opens browser's native print dialog
  • Can save as PDF from the print dialog
  • Optimized print layout hides navigation elements
  • Score badges preserve colors

Use case: Quick printouts for meetings or filing

HTML Export

  • Downloads a standalone HTML file
  • Embedded CSS for consistent appearance
  • All evaluation data included:
    • Domain and score with color-coded badge
    • Confidence level
    • Category and bias indicator
    • LLM reasoning
    • Evidence cited (parsed and formatted)
    • Evidence pack with source links
    • Model information and consensus status
    • Cache metadata

Filename format: {domain}_source_eval_{YYYYMMDD_HHMMSS}.html

Use case: Sharing evaluations via email, archival

Markdown Export

  • Downloads a formatted .md file
  • Tables for structured data (summary, models, cache info)
  • Blockquotes for LLM reasoning
  • Lists for evidence

Filename format: {domain}_source_eval_{YYYYMMDD_HHMMSS}.md

Use case: Documentation, version control, wikis

JSON Export

  • Downloads structured JSON data
  • Complete CachedScore object
  • Parsed JSON fields (evidenceCited, evidencePack)
  • Export metadata included
  • Pretty-printed with 2-space indentation

Filename format: {domain}_source_eval_{YYYYMMDD_HHMMSS}.json

Use case: Data analysis, API integration, backups

Exported Data Fields

All export formats include:

 Field  Description
 Domain  The evaluated source domain
 Score  Reliability score (0-100)
 Confidence  LLM confidence in the evaluation
 Category  Source category (news, academic, etc.)
 Bias  Detected bias indicator
 Reasoning  LLM's reasoning for the score
 Evidence Cited  Specific facts the LLM used
 Evidence Pack  Source URLs consulted
 Primary Model  Main LLM used for evaluation
 Secondary Model  Verification LLM (if multi-model enabled)
 Consensus  Whether models agreed
 Evaluated At  When the evaluation was performed
 Expires At  When the cached evaluation expires

Known Limitations

  1. Print functionality depends on browser's print dialog capabilities
  2. Score badge colors in print may vary based on browser's color handling
  3. Very long URLs in evidence pack may wrap in print view
  4. Export is limited to the currently selected entry (no bulk export)

Related Documentation