Fork it - Mess with it - Change the models - Change the prompts c: - Create your own council!

Multi-Model Blind Peer Review System

Submit your documents to multiple AI models for comprehensive, unbiased peer review. Each reviewer brings a unique perspective, ensuring thorough analysis from technical rigor to ethical considerations. Then Claude Opus 4.5 synthesizes each response into a meta-review.

• Blind Review: Removes author identity to prevent bias
• Diverse Perspectives: Uses different AI models with distinct personas
• Multi-dimensional Analysis: Covers methodology, logic, ethics, feasibility, and more
• Automated Synthesis: Consolidates feedback into actionable summaries with verdicts
• Fallback Handling: Automatically uses backup models if primary ones fail

6 diverse reviewers covering traditional peer review perspectives:

12 specialized reviewers (original 6 + 6 philosophical/systems reviewers):

pip install -r requirements.txt

Or install manually:

pip install requests python-dotenv

Get your API key from OpenRouter:

1. Sign in or create an account
2. Navigate to the Keys section
3. Create a new API key

Then set it up using one of these methods:

Option A: Environment Variable (Recommended for one-time use)

# Linux/Mac
export OPENROUTER_API_KEY='your-key-here'

# Windows (PowerShell)
$env:OPENROUTER_API_KEY = 'your-key-here'

# Windows (CMD)
set OPENROUTER_API_KEY=your-key-here

Option B: .env File (Recommended for persistent use)

Create a .env file in the project directory:

OPENROUTER_API_KEY=your-key-here

Standard (6 reviewers):

python trial_by_hex.py ./your_document.md ./review_output.md

Enhanced (12 reviewers):

python trial_by_hex_plus.py ./your_document.md ./review_output.md

The review output includes:

1. Synthesized Review - Consensus-based summary with:

   • Critical issues (4+ reviewers agree)
   • Significant concerns (2-3 reviewers)
   • Individual considerations
   • Strengths
   • Overall verdict (PASS / REVISE / REJECT)

2. Individual Reviews - Full feedback from each reviewer

## High Consensus (4+ reviewers agree)
[Issues most reviewers flagged]

## Moderate Consensus (2-3 reviewers)
[Issues some reviewers noted]

## Minority Concerns (1 reviewer, but substantive)
[Individual concerns worth considering]

## Strengths (what reviewers praised)
[Positive feedback]

## VERDICT
PASS: Ready for publication with minor edits
REVISE: Needs significant revision, re-review recommended
REJECT: Fundamental issues need addressing

## CRITICAL ISSUES (Consensus across multiple reviewers)
[Issues flagged by 4+ reviewers - these are blockers]

## SIGNIFICANT CONCERNS (2-3 reviewers)
[Important issues worth addressing]

## CONSIDERATIONS (Single reviewer, but substantive)
[Individual concerns that deserve thought]

## STRENGTHS (What reviewers praised)
[Positive consensus]

## LOGICAL/FORMAL ISSUES
[From the Logical Consistency Reviewer - Godelian concerns, self-reference issues]

## LINGUISTIC CLARITY
[From the Semantic Analyst - language bewitchments, definition problems]

## ETHICAL CONSIDERATIONS
[From the Ethical Alignment Sentinel - bias, impact, safety]

## IMPLEMENTATION FEASIBILITY
[From the Systems Architect - can it be built?]

## INTERDISCIPLINARY CONNECTIONS
[From the Interdisciplinary Catalyst - missed connections, synthesis opportunities]

## STEEL MANNED VERSION
[From the Steel Man Advocate - the strongest form of the argument]

## VERDICT
**Technical Quality:** ⭐⭐⭐⭐☆
**Logical Coherence:** ⭐⭐⭐⭐☆
**Ethical Alignment:** ⭐⭐⭐⭐⭐
**Feasibility:** ⭐⭐⭐☆☆
**Novelty:** ⭐⭐⭐⭐☆

**Overall:** REVISE

**Priority Actions:**
1. [Most important fix]
2. [Second priority]
3. [Third priority]

• Python 3.7+
• requests library
• python-dotenv library (optional, for .env file support)
• OpenRouter API key

Trial by Hex uses multiple AI models via OpenRouter. Costs vary by model:

• Standard (6 reviewers): ~$0.10-0.50 per review (varies by document length)
• Plus (12 reviewers): ~$0.20-1.00 per review (uses more premium models)

The scripts include fallback to cheaper/free models if primary models fail.

Any text-based document:

• Markdown (.md)
• Plain text (.txt)
• Academic papers
• Research proposals
• Technical documentation
• Essays and articles

Trial-by-Hex/
├── trial_by_hex.py      # Standard 6-reviewer version
├── trial_by_hex_plus.py # Enhanced 12-reviewer version
├── requirements.txt     # Python dependencies
├── .env.example         # Example environment file
├── README.md            # This file
└── LICENSE              # MIT License

Make sure you've set your API key using one of the methods above. If using a .env file, ensure it's in the same directory where you run the script.

Run: pip install requests

The scripts automatically fall back to cheaper models. If issues persist, check your OpenRouter account for rate limits or credit balance.

The scripts use UTF-8 encoding. Ensure your input documents are saved with UTF-8 encoding.

License Change: 2nd December 2025

As of 2nd December 2025, Trial by Hex is licensed under the GNU General Public License v3.0.

If you downloaded Trial by Hex prior to this date (27th November - 1st December 2025), that copy remains covered under the MIT License that was included at the time. All I ask is that you follow the MIT agreement and credit me.

All downloads from 2nd December 2025 onwards are licensed under GPLv3. This means any derivative works must also be open source under GPLv3.

See the LICENSE file for full terms.

• Eden Eldith
• Claude Opus 4.5

Contributions welcome! Feel free to:

• Add new reviewer personas
• Improve synthesis prompts
• Add support for additional output formats
• Enhance error handling

"Six reviewers stand in judgment."

"None know whose work they judge."

"Let the ideas speak for themselves."

Made by a human and AI in collaboration

Copyright (C) 2025 Eden_Eldith (P.C. O'Brien) c: