Samarth-S-Shetty/ticket-triage-agent

AI-powered ticket triage system using LLM extraction + embedding search FastAPI backend • Dockerized • Production ready

What's novel

AI-powered ticket triage system using LLM extraction + embedding search FastAPI backend • Dockerized • Production ready

Code Analysis

9 files read · 3 rounds

A FastAPI service that uses an LLM to extract structured fields from support tickets and embedding-based cosine similarity search against a static JSON knowledge base to classify tickets as known or new issues with recommended next actions.

Strengths

Clean separation of concerns across app/agent/kb modules, genuine implementation of embedding search with disk caching and cosine similarity, and sensible fallback chains (LLM→mock, embeddings→keyword) that make the system functional without an API key.

Weaknesses

Tests are minimal (3 basic tests, no mocking, no edge cases for search logic), the rate limiter is a naive global dict with no cleanup that blocks all endpoints, and the README's 'production-ready' claim is overstated given the lack of async handling, structured logging, and comprehensive testing.

Score Breakdown

Innovation

4 (25%)

Craft

57 (35%)

Traction

6 (15%)

Scope

59 (25%)

Signal breakdown

Innovation

Not Fork+1

Code Novelty+1

Concept Novelty+1

Craft

Ci-1

Tests+5

Polish+0

Releases+0

Has License+0

Code Quality+19

Readme Quality+15

Recent Activity+7

Structure Quality+5

Commit Consistency+2

Has Dependency Mgmt+5

Traction

Forks+0

Stars+6

Hn Points+0

Watchers+0

Early Traction+0

Devto Reactions+0

Community Contribs+0

Scope

Commits+7

Languages+5

Subsystems+5

Bloat Penalty+0

Completeness+6

Contributors+5

Authored Files+8

Readme Code Match+3

Architecture Depth+3

Implementation Depth+8

Evidence

Commits

Contributors

Files

Active weeks

TestsCI/CDREADMELicenseContributing

Repository

Language

Python

Stars

Forks

License

—