BroCheck — The Open-Source AI Roaster for Playlists, Code & GitHub Profiles

Nagaraj Nikhil RB — Sat, 03 Oct 2026 16:28:59 +0000

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built

We’ve all got that one friend whose taste and habits desperately need a reality check:

The friend who puts Nickelback, Baby Shark, and Crazy Frog on the same 3 AM roadtrip playlist.
The developer friend who names variables temp_var_2_final_really_final and writes 600-line if/else ladders.
The friend whose GitHub profile has 0 green commit squares this year but their bio screams "10x Senior Full-Stack Architect".
The friend whose excuses for being 2 hours late to a group hangout defy the laws of physics.

To lovingly solve this problem, I built BroCheck — a lightweight, multi-mode terminal CLI powered by Open-Source AI that delivers customized, hilarious roasts on whatever you throw at it.

Key Features:

Universal Music Playlist Roaster: Drop any playlist link (Spotify, Apple Music, YouTube Music, SoundCloud), local .txt/.m3u files, or raw artist lists.
Code Review Roaster: Analyses any code file (.py, .js, .ts, .cpp, etc.) for architectural and stylistic sins.
GitHub Profile Roaster: Scrapes public GitHub stats, repos, and bios without requiring API keys, roasting developer habits.
Social Bio & Status Roaster: Dismantles fake-deep gym quotes and corporate LinkedIn buzzwords.
Chat Excuse Roaster: Calls out group chat lies and late arrival excuses.
3 Heat Levels: Mild (playful banter), Spicy (sharp satire), and Nuclear (pure comedic devastation).
4 Personas: Best Friend (casual slang), Gordon Ramsay ("IT'S RAW!"), Tech Bro VC (negative alpha / zero ROI), and Disappointed Parent ("Sharma-ji's son would never").

Demo

Here is BroCheck in action across different modes and personas:

1. Roasting a Friend's Roadtrip Playlist (Best Friend Persona, Spicy Heat)

$ python brocheck.py --mode playlist --input "Nickelback, Baby Shark, Imagine Dragons, Crazy Frog" --heat spicy --persona bestfriend

╭──────── BROCHECK RESULTS: Best Friend (Loving Savage) (Spicy) ─────────╮
│                                                                        │
│  Bro... honestly, who hurt you? What is this actual abomination?       │
│                                                                        │
│  I just looked at 'Crazy Frog & Baby Shark' and my brain cells         │
│  literally filed for unemployment. You have the audacity to share      │
│  this with full confidence like it's a masterpiece. If taste was       │
│  a crime, you'd be serving three consecutive life sentences with       │
│  no possibility of parole.                                             │
│                                                                        │
│  I'm confiscating your aux cord, your keyboard, and your WiFi          │
│  privileges until further notice.                                      │
│                                                                        │
│  FINAL VERDICT: Unhinged, certified criminal offense. Please seek       │
│  immediate help.                                                       │
│                                                                        │
╰────────────────────────────────────────────────────────────────────────╯

2. Roasting Spaghetti Code (Gordon Ramsay Persona, Nuclear Heat)

$ python brocheck.py --mode code --file demo_samples/bad_code.py --heat nuclear --persona gordon-ramsay

╭──── BROCHECK RESULTS: Gordon Ramsay (Kitchen Screamer) (Nuclear) ────╮
│                                                                      │
│  LISTEN TO ME! LOOK AT THIS! THIS IS AN ABSOLUTE DISASTER!           │
│                                                                      │
│  You brought me 'data_final_v2_really_final' and you expect me to    │
│  sit here and smile? It is RAW! It has NO FLAVOR, NO PASSION, AND    │
│  ZERO QUALITY CONTROL! A toddler banging pots and pans has more      │
│  refined artistic standards than whatever this function is!          │
│                                                                      │
│  SHUT IT DOWN! Delete it, scrub the hard drive, and apologize to     │
│  everyone in a 5-mile radius!                                        │
│                                                                      │
│  FINAL VERDICT: An absolute culinary and acoustic catastrophe.       │
│  Get out of my kitchen!                                              │
│                                                                      │
╰──────────────────────────────────────────────────────────────────────╯

Code

The entire codebase is open-source under the MIT license:

NICKSTER0506 / Bro_Check

BroCheck — a lightweight, multi-mode terminal CLI powered by Open-Source AI that delivers customized, hilarious roasts on whatever you throw at it.

BroCheck

The Open-Source AI Roaster CLI
Built for the Hacktoberfest 2026 (HF26) DEV Challenge: Build for a Friend

BroCheck is a zero-RAM, multi-mode AI roaster CLI that hilariously critiques your friend's music playlists, spaghetti code, barren GitHub profiles, cringey bios, and transparent excuses using Open-Source AI.

Features

Universal Music Roaster: Drop any playlist link (Spotify, Apple Music, YouTube Music, SoundCloud), upload a .txt/.m3u file, or paste raw song lists.
Code Review Roaster: Point to any .py, .js, .ts, .cpp file for a brutally honest critique of variable names and 6-level nested loops.
GitHub Profile Roaster: Enter any GitHub username to roast empty commit heatmaps, abandoned test repos, and buzzword-packed bios.
Social Bio & Status Roaster: Savage fake-deep gym quotes and LinkedIn buzzwords.
Excuse & Chat Roaster: Roast terrible excuses for being late to group hangs.
Adjustable Heat…

View on GitHub

Project Structure:

BroCheck/
├── brocheck.py                 # Main CLI with ASCII banner & interactive wizard
├── engine/
│   ├── llm_client.py           # Open-weight AI engine (0 MB RAM overhead)
│   ├── prompts.py              # 4 Personas & 3 Heat levels
│   ├── playlist_parser.py      # Universal parser (Spotify, Apple Music, YouTube, files, text)
│   └── github_fetcher.py       # Public GitHub scraper
├── demo_samples/
│   ├── bad_code.py             # Sample messy code for testing
│   └── friend_playlist.txt     # Sample playlist for testing
└── requirements.txt            # Minimal dependencies (rich, requests)

How I Built It

BroCheck was built using Python and centered around open-source AI:

flowchart TD
    CLI[User Terminal / Interactive Wizard] --> Parser{Universal Input Parser}
    Parser -->|Spotify / Apple / YT Link| LinkExtractor[oEmbed & Metadata Scraper]
    Parser -->|.txt / .m3u File| FileHandler[Local File Reader]
    Parser -->|GitHub Handle| GHAPI[GitHub Public API]

    LinkExtractor & FileHandler & GHAPI --> PromptEngine[Prompt & Persona Builder]
    PromptEngine --> OpenEngine[Open-Weight AI Engine\nLlama-3.2 / Qwen-2.5 / Mistral]
    OpenEngine --> RichUI[Rich Terminal UI & Verdict Card]

Open-Weight Models: Powered by meta-llama/Llama-3.2-3B-Instruct, Qwen/Qwen2.5-Coder, and mistralai/Mistral-7B.
Universal Playlist Scraper: Leverages public oEmbed and metadata endpoints to parse links from Spotify, Apple Music, and YouTube without requiring private API tokens.
Public GitHub Fetcher: Retrieves repository names, commit frequencies, and bios directly from GitHub's public API.
Rich Terminal Engine: Provides an interactive setup wizard, spinners, result panels, and customizable terminal themes using Python rich.

Why Does Open Innovation Matter?

This project was developed on a machine with limited RAM (6GB), where running heavy local LLMs would cause system lag and memory thrashing.

Open innovation solved this completely:

Freedom of Open Weights: We utilize open-weight models (Llama-3.2, Qwen-2.5, Mistral) instead of being locked into closed, proprietary platforms with unpredictable pricing and opaque modifications.
0 MB RAM Footprint: By pairing a lightweight Python client harness with open serverless inference, anyone with budget or low-spec hardware can run and build AI tools with sub-second response times.
Friend Data Privacy: Personal group chat texts, private code snippets, and quirky playlist links are not fed into closed commercial model training loops.
Uncensored Creative Writing: Closed AI APIs often over-moderate playful humor and satirical character prompts. Open-weight models give developers full control over system instructions, enabling fun personas like Gordon Ramsay screaming at unseasoned code.

My Agent Session

This project was designed, planned, scaffolded, and tested with DevRelay and Antigravity.

Happy Hacktoberfest 2026! Protect your code, protect your aux cord.

DEV Community: Nagaraj Nikhil RB

Check this out :)