Slangify

Make your AI coding assistant speak your language.

---

Claude / Codex / Cursor normally:

"I found an issue with the authentication flow."

With Slangify:

• 🇮🇳 Mumbai

  "Bhai auth flow mein lafda hai. Token validation baad mein ho raha hai, isko pehle shift karte hain."

• 😂 Gen Z

  "bro the auth flow is cooked 💀 token validation is happening way too late. fixing it now."

• 🇬🇧 London

  "Bit of a mess in the auth flow, mate. Token validation is running after the request is processed."

• 👔 Corporate

  "We should revisit the authentication strategy to optimize our security posture and align deliverables."

• 🌾 Haryanvi

  "Bhai auth ka flow gadbad se. Token baad mein check kar rya hai, pehle karna padega."

• 🗽 NYC

  "Deadass, the auth flow is outta pocket my guy. Let's lock this down right now, forget about it."

• 🦘 Aussie

  "Crikey mate, the auth flow is completely cactus. Let's get that sorted, she'll be right."

• 🌶️ Brutally Honest (Code Roast)

  "Bro this function is 180 lines long. At this point it's not a function, it's a lifestyle choice."

---

⚡ Quick Start (10 Seconds)

1. Install integrations into your coding agents

npx slangify install

Slangify

Detected coding agents:
✓ OpenAI Codex
✓ Claude Code
✓ Cursor

Installed Slangify integrations into:
- OpenAI Codex: AGENTS.md
- Claude Code: CLAUDE.md
- Cursor: .cursor/rules/slangify.mdc

2. Choose your vibe

slangify use mumbai

3. Open your agent and code normally!

Open Claude Code, Cursor, or Codex and start coding.

Developer: Fix the missing dependency error in this repo.

AI Agent:  Bhai dependency missing hai. Run:
           npm install express
           Isko daalo, phir build ek number pass ho jayega!

Note

No proxy. No copy-pasting. No extra apps running in the background.
Slangify installs directly into the agent's native instructions. The AI agent itself produces the authentic responses natively.

---

🛡️ The Cardinal Rule: HOW vs. WHAT

Slangify changes:

HOW your AI communicates

It NEVER changes:

WHAT your AI does

❌ Never modifies:
   - Source code syntax
   - Terminal commands
   - File paths
   - Variable names
   - API endpoints & schemas
   - SQL queries
   - JSON / YAML configs
   - Stack traces & error messages

For example:

"Bhai dependency missing hai. Run npm install express."

The terminal command npm install express remains 100% exact and copy-pasteable. Your tests, types, and logic remain pristine.

---

🎭 Available Personalities

Run slangify list to explore all styles:

Available Slangify styles:

Regional
  mumbai           Casual, street-smart Mumbai Hinglish (Bhai, lafda, jugaad, sorted)
  delhi            Bold, swagger-filled Delhi NCR Hinglish (Yaar, scene, chill maar)
  haryanvi         Direct, grounded Haryanvi style (Bhai, rya se, ghana, lath gaad diya)
  up-bihar         Flavorful eastern UP & Bihari style (Bhaiya, bawaal, gardaa uda diye)
  london           Casual, witty British & London vernacular (Mate, proper, sorted, innit)
  nyc              Fast-paced NYC street vibe (Deadass, facts, outta pocket, my guy)
  aussie           Laid-back Australian English (G'day mate, no worries, cactus)

Internet
  gen-z            Internet-native Gen Z slang (cooked 💀, no cap, fr fr, bet, lowkey)
  brainrot         Peak Gen-Alpha memes (Skibidi, rizzler, fanum tax, ohio, sigma)

Professional
  corporate        Satirical ultra-corporate buzzwords (Circle back, synergize, bandwidth)
  brutally-honest  Unfiltered code-review roasts with surgical fixes

---

🎚️ Slang Intensity (Levels 1–5)

Control how much flavor the AI uses with --level:

# Level 1: Subtle (mostly standard English with a slight conversational nod)
slangify use mumbai --level 1
# "Found a small bug here, bhai. The token expired."

# Level 3: Balanced (natural, everyday teammate tone - Default)
slangify use mumbai --level 3
# "Bhai auth middleware mein lafda hai. Token baad mein check kar raha hai."

# Level 5: Full Flavor (maximum cultural expression & energy)
slangify use mumbai --level 5
# "Arre bawa, ye function dekh ke production ki vat lag jayegi! Apun abhi jugaad karke sorted karta hai!"

---

🎪 Viral Feature: Compare Personalities

Curious how each personality would tackle the same code critique? Run:

slangify compare "This function has too many responsibilities."

NORMAL (DEFAULT AI)
This function has too many responsibilities.

MUMBAI
Bhai ek function mein pura department ghusa diya hai. Isko alag-alag karo.

GEN Z
bro this function is doing everyone's job 💀 pure unhinged vibes fr fr.

CORPORATE
We should revisit the distribution of responsibilities to optimize bandwidth and leverage core competencies.

BRUTALLY HONEST
Bro this function is 180 lines long. At this point it's not a function, it's a lifestyle choice.

NYC
Yeah, this function is doing way too much, deadass. Let's break it down, my guy.

LONDON
Bit of an absolute dog's dinner this function, mate. Doing far too much at once.

---

💻 CLI Commands Reference

Command	Description
slangify install	Detects installed agents and configures Slangify integrations
slangify uninstall	Cleanly removes Slangify and restores original configuration backups
slangify list	Lists all available personality packs
slangify use <style>	Activates a personality (--level 1-5, --global, --agent <name>)
slangify current	Shows currently active style, scope, intensity, and active agents
slangify on	Turns Slangify ON
slangify off	Turns Slangify OFF (returns AI to standard default tone)
slangify status	Detailed overview of config files and agent integration paths
slangify compare [text]	Side-by-side personality showcase

---

📁 Global vs. Project Configuration

Slangify supports scoping at both the project and global level:

• Global (~/.slangify/config.json): Applies to all coding projects on your machine.

  slangify use gen-z --global

• Project (.slangify/config.json): Scoped to the current repository. Project settings automatically override global settings!

  slangify use corporate # In work repo

This allows:

• Work Repo ➡️ Corporate
• Side Project A ➡️ Mumbai
• Open Source Project B ➡️ Brutally Honest

---

🔌 Architecture & Supported Agents

Slangify uses a plugin-first, local-first architecture. It communicates directly through native instruction standards:

Developer
    ↓
AI Coding Agent (Claude Code / Cursor / Codex)
    ↓
Native Instructions (CLAUDE.md / .cursor/rules / AGENTS.md)
    ↓
AI Response in your favorite slang

Agent	Integration Mechanism
Claude Code	Root CLAUDE.md / .claude/rules/slangify.md (Project) & ~/.claude/CLAUDE.md (Global)
Cursor	.cursor/rules/slangify.mdc (alwaysApply: true) & .cursorrules
OpenAI Codex	AGENTS.md (Modern Multi-Agent Standard) & .codex/instructions.md

Non-Destructive Safety

Slangify wraps its rules inside demarcated boundary markers:

<!-- SLANGIFY:START -->
...
<!-- SLANGIFY:END -->

If you already have rules in CLAUDE.md or AGENTS.md, Slangify preserves them completely and creates automated .slangify-backup files before modifying anything.

---

🤝 Adding Your Own Slang (Community Packs)

Adding a personality requires zero code! Just drop a folder under packs/:

packs/
  my-city/
    manifest.json
    instructions.md
    examples.md

See CONTRIBUTING.md for the complete 5-minute guide.

---

📄 License

MIT © Slangify Contributors