Somm
Powered by Gemini 3NEW

AI Code
Evaluation
with Sommelier
Sophistication

Six specialized AI agents analyze your repositories from every angle— structure, quality, security, and innovation.

✨ Free for open source🔒 No signup required

YOUR CODE

See What You Get

A comprehensive evaluation report with actionable insights

Repository Evaluation

ai/nanoid

Final Score

97/100

Legendary— Exceptional quality codebase
Marcel

Marcel

96

TOP
Isabella

Isabella

98

Heinrich

Heinrich

96

Sofia

Sofia

96

TOP
Laurent

Laurent

98

Jean-Pierre

Jean-Pierre

97

* Real evaluation of ai/nanoid repository

How It Works

Like a master sommelier orchestrates a tasting panel, our AI agents analyze your code in parallel.

1

Submit Repository

Connect your GitHub repository or paste any public repo URL

Public or private repos
2

Parallel Analysis

Six specialized sommeliers evaluate simultaneously

marcel
isabella
heinrich
sofia
laurent
+1
3

Get Verdict

Jean-Pierre synthesizes all findings into actionable insights

Jean-Pierre

Jean-Pierre

Grand Sommelier

Features

Everything you need for comprehensive code evaluation

LIVE

Real-time Streaming

Watch evaluations unfold live with Server-Sent Events streaming. See each sommelier's analysis as it happens.

Live updates• No page refresh

PDF Reports

Export professional tasting notes as beautifully formatted PDF reports.

One-Click Share

Share results with your team instantly via link or social media.

4 Evaluation Modes

Choose the perfect evaluation style for your needs.

Basic🏆 HackathonAcademicCustom

Ready to evaluate your code?

Get detailed insights from six AI sommelier agents in minutes.

Free for open source No signup required
Try It Now Takes less than 2 minutes

Evaluation Modes

Choose the criteria that fits your needs

🏆

Hackathon Mode

RECOMMENDED

Gemini 3 judging criteria • Best for hackathon submissions

Gemini judging focusInnovation & ImpactTechnical creativityDemo-ready polish
Use This Mode
📋

Basic

Balanced

General code review

Code structure
Best practices
Documentation

Best for: Everyday projects

🎓

Academic

Novelty

Research-focused evaluation

Methodology
Reproducibility
Novelty

Best for: Research projects

⚙️

Custom

Flexible

Define your own criteria

Your criteria
Custom weights
Flexible scoring

Best for: Special requirements

Meet Your Sommeliers

Six AI experts, each with a unique perspective on code quality

Marcel
Isabella
Heinrich
Sofia
Laurent
Jean-Pierre
Structure & Metrics

Marcel

Cellar Master

A well-organized cellar reveals the soul of its maker. I examine the architecture, file organization, and code metrics to understand the structural integrity of your codebase. 

Scoring System

From legendary cellars to house wines — every codebase has its place

Legendary
95-100
Grand Cru
90-94
Premier Cru
85-89
Village
80-84
Table Wine
70-79
House Wine
60-69
Corked
<60
Exceptional
Good
Basic

Ready to taste your code?

Get comprehensive insights from six AI sommeliers in minutes.

Start Your Evaluation
Somm

© 2026 Somm.dev — AI Code Sommelier

GitHub