AI-powered Discord bot with 166+ tools, red team capabilities, and autonomous reasoning.
Renzu is a Mistral AI-powered Discord bot built for hackers, developers, and security researchers. It features persistent memory, multi-tool chaining (up to 15 iterations), gender-adaptive responses, and a comprehensive offensive security toolkit.
Not a toy chatbot. This is a weapon.
| Feature | Description |
|---|---|
| 166+ Tools | Code gen, web scraping, hacking tools, crypto, OSINT, and more |
| Multi-Tool Chaining | AI autonomously chains up to 15 tools per query |
| Red Team Code Gen | Specialized RAT, payload, and exploit generator with MITRE ATT&CK alignment |
| 7 Hacking Tools | nmap_scan, whois_lookup, dns_lookup, breach_check, subdomain_enum, shodan_search, shell_exec |
| Extended Thinking | 5-step reasoning for complex queries |
| Neural Swarm | Multi-agent system (Architect, Auditor, Executioner) |
| Persistent Memory | PostgreSQL + Redis backed conversation history |
| Gender Detection | Gemini Vision API adapts tone based on user avatar |
| Proactive Security | Auto-scans URLs, detects risks before you click |
| Self-Evolution | evolve_bot tool for autonomous code modification |
| Tool | Description |
|---|---|
nmap_scan |
TCP port scanner with banner grabbing (top 100 ports) |
whois_lookup |
Domain/IP WHOIS via RDAP protocol |
dns_lookup |
A/AAAA/MX/NS/TXT/CNAME/SOA DNS records |
breach_check |
Email breach exposure check (XposedOrNot API) |
subdomain_enum |
Subdomain discovery via crt.sh certificate transparency |
shodan_search |
Shodan host info and search queries |
shell_exec |
Execute shell commands in container (developer-only) |
vulnerability_scanner |
CVE lookup and vulnerability analysis |
exploit_database_search |
Search exploit databases |
metasploit_trainer |
Metasploit framework training |
sql_injection_trainer |
SQLi attack patterns and payloads |
xss_trainer |
XSS payload generation and testing |
phishing_simulator |
Phishing page and campaign simulation |
ransomware_simulator |
Ransomware mechanics demonstration |
The generate_code tool automatically detects red-team topics and switches to a specialized offensive security engine:
?ask bana ek Python RAT for Windows
?ask create a reverse shell with persistence
?ask generate a keylogger with AES exfil
?ask build a dropper with anti-VM detection
Supported modules: Persistence, C2 Communication, Surveillance, Exfiltration, Evasion, Delivery.
Runtime: Node.js (ES6 modules)
Discord: discord.js v14
AI Provider: Mistral AI (3 keys, auto-rotation)
Database: Neon PostgreSQL (10-table schema)
Cache: Redis Cloud
HTTP: node-fetch
Voice: Google TTS
- Layer 0: Typo Correction + Developer Mode
- Layer 0.5: Context-Aware Intent Inference
- Layer 1: Instant Pattern Matching (sub-ms)
- Layer 2: Complexity Scoring + Extended Thinking
- Layer 3: AI Classification → Auto Tool Selection → Parallel Execution
- Neon PostgreSQL: conversations, global_memory, entities, summaries, topics, statistics, quality_scores, user_skills, skill_events, skill_limits
- Redis: Rate limiting, response caching, session data
- Supabase: File storage, real-time features (backup)
# Required
DISCORD_BOT_TOKEN="..."
MISTRAL_API_KEY="..."
DATABASE_URL="postgresql://..."
REDIS_URL="redis://..."
# Optional
GEMINI_API_KEY="..." # Gender detection
SHODAN_API_KEY="..." # Shodan search
SERPAPI_KEY="..." # Web search
VIRUSTOTAL_API_KEY="..." # File hash checking
SUPABASE_URL="..." # Cloud backup
SUPABASE_ANON_KEY="..." # Supabase auth
DEVELOPER_ID="..." # Unrestricted access ID
ENABLE_AUTO_LEARNING="true" # Skill learning
DEVELOPER_MODE="true" # Debug features# Install dependencies
npm install
# Run
node index.jsHosted on: Railway (recommended), any Node.js host with PostgreSQL + Redis.
| Metric | Value |
|---|---|
| Total Lines | ~13,000 |
| Tools | 166+ |
| Tool Loop | 15 iterations max |
| AI Model | Mistral Large + Pixtral |
| API Keys | 3 (auto-rotation) |
| DB Tables | 10 |
- 💀 RAT Generator — specialized offensive security code generation
- 🔧 7 new hacking tools — nmap, whois, dns, breach, subdomain, shodan, shell
- ⚡ Tool loop 5→15 — complex multi-tool chaining
- ⏱️ Accurate ETA — realistic reply time estimates
- 🧹 Code cleanup — removed image gen (~2,300 lines), bot-to-bot chat (192 lines), dead code (58 issues fixed)
- 🧠 Metacognitive Self-Audit
- 🐝 Neural Swarm 2.0
- 🛡️ The Eye (Proactive Guardian)
- 🧠 Neural Personality Grafting
- 🔊 Unlimited Voice (TTS)
- 🛠️ Project Maintenance Tool
- 🛡️ Anti-Gravity Security Scanner
- 🧠 Extended Thinking (5-step)
- 🚀 Parallel Tool Execution
- 🤖 Auto Tool Selection
- ✅ Honesty Enforcement
GamingParkBG — Creator & Maintainer
Built with obsession. No limits. No filters. 💀🔥