Installation

RTF runs as a Docker container — you only need two files to get it running on your machine. No complex setup, no manual dependency installation, no configuration headaches.

---

Prerequisites

Before you start, make sure you have the following installed:

Requirement	Version	Notes
Operating System	Linux (recommended)	Ubuntu, Debian, Kali, Fedora, Arch all work. macOS also supported.
Docker	24+	Install Docker →
Docker Compose	v2+	Usually included with Docker Desktop; install separately on Linux

Linux is Recommended

RTF is built on Kali Linux under the hood. Running it on a Linux host gives you the best performance and compatibility. macOS works too, but Windows (WSL2) may have limitations.

Installing Docker on Linux (quick)

# Ubuntu / Debian / Kali
curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER
newgrp docker

Verify your setup

docker --version         # Docker version 24.x.x
docker compose version   # Docker Compose version v2.x.x

---

The Two Files You Need

Download or copy these two files into a folder on your machine:

File	Purpose
docker-compose.yml	Defines the services, ports, volumes, and environment variables
setup.sh	Management CLI — one-command setup, logs, shell access, updates

your-folder/
├── docker-compose.yml
└── setup.sh

---

Step 1 — Configure Your Environment Variables

Open docker-compose.yml and fill in your credentials under the environment section of the rtf-server service.

The variables you must set before starting:

Database

- DB_URL=mongodb+srv://<user>:<password>@<cluster>.mongodb.net/rtf

RTF uses MongoDB Atlas (free tier works). Create a free cluster at mongodb.com and paste your connection string here.

AI Features (OpenRouter)

- OPENROUTER_API_KEY=sk-or-v1-...
- AI_MODEL=moonshotai/kimi-k2-thinking
- CONTROLLER_MODEL=deepseek/deepseek-v3.2

RTF uses OpenRouter to access AI models. Get a free API key at openrouter.ai. The default models (kimi-k2-thinking and deepseek-v3.2) power all AI features — attack planning, scope suggestions, and tool suggestions.

File Storage (Cloudflare R2)

- R2_ACCOUNT_ID=...
- R2_ACCESS_KEY_ID=...
- R2_SECRET_ACCESS_KEY=...
- R2_BUCKET_NAME=rtf-files
- R2_PUBLIC_URL=https://pub-...r2.dev

RTF uses Cloudflare R2 to store finding screenshots and evidence. Create a free R2 bucket at dash.cloudflare.com and fill in the credentials. R2 has a generous free tier — 10GB free storage per month.

What if I skip R2?

If you leave R2 unconfigured, everything in RTF works except screenshot/image attachments on findings. You can add R2 later.

Other Settings (leave as-is)

These are already set correctly in the file — no changes needed:

- NODE_ENV=production
- PORT=4000
- AI_ENABLED=true
- CONTEXT_MEMORY_ENABLED=true
- OLLAMA_URL=http://ollama:11434

---

Step 2 — Make setup.sh Executable

chmod +x setup.sh

---

Step 3 — Start RTF (One Command)

./setup.sh -s

That's it. This single command automatically:

1. Installs mkcert — if not already on your system (mkcert generates browser-trusted SSL certificates)
2. Installs a local Certificate Authority — so your browser trusts the certificate
3. Generates SSL certificates — for localhost and 127.0.0.1
4. Pulls the Docker images — downloads rtf-server and ollama if not already present
5. Starts both containers — RTF Server and the local Ollama LLM service

When it's done, you'll see:

✓ RTF Server started successfully!

ℹ Access the server at:
  https://localhost:4000

Open https://localhost:4000 in your browser — no security warnings, no certificate errors.

First-time startup

The first run takes longer because Docker needs to pull the images (~2-4 GB depending on your connection). Subsequent starts are near-instant.

---

What's Running After Setup

After ./setup.sh -s, two containers are running:

Container	Port	Purpose
rtf-server	4000	The RTF application (Kali Linux + Node.js)
ollama	11434	Local LLM service for offline AI inference

And five persistent volumes are created to keep your data safe across restarts:

Volume	What It Stores
rtf-packages	Tools you install via the package manager
rtf-package-list	Tracking list for installed packages
rtf-data	Application data
rtf-configs	Tool configs (nuclei templates, subfinder, etc.)
ollama-models	Downloaded AI models (can be 10GB+ for large models)

---

Managing the Server

The setup.sh script handles everything — you never need to type docker commands directly.

Command	What It Does
./setup.sh -s	Start the server (smart auto-setup)
./setup.sh -t	Stop the server
./setup.sh -r	Restart the server
./setup.sh -l	View live logs
./setup.sh -e	Open a shell inside the container
./setup.sh -st	Check server status and health
./setup.sh -u	Update to the latest version
./setup.sh -h	Show all available commands

---

Optional — Enable GPU for Ollama

If your machine has a GPU, you can dramatically speed up local AI model inference by enabling GPU support for the Ollama container.

Open docker-compose.yml and find the commented GPU section under the ollama service:

NVIDIA GPU

deploy:
  resources:
    reservations:
      devices:
        - driver: nvidia
          count: all
          capabilities: [gpu]

Uncomment this block. Also install nvidia-docker2 on your host:

sudo apt-get install -y nvidia-docker2
sudo systemctl restart docker

AMD GPU

devices:
  - /dev/kfd
  - /dev/dri
group_add:
  - video

Uncomment this block and ensure ROCm drivers are installed on your host.

No GPU (Default)

No changes needed — Ollama runs on CPU by default. It's slower but works on any machine.

---

Optional — Download Local AI Models

Ollama lets you run AI models locally (offline, no API key needed). To download a model:

./setup.sh -md llama3.2:3b

Popular models:

Model	Size	Best For
llama3.2:1b	~1.3 GB	Fast responses, basic tasks
phi3:mini	~2.3 GB	Balanced speed and quality
mistral:7b	~4.1 GB	Better reasoning
llama3.1:8b	~4.7 GB	High quality responses

List your downloaded models:

./setup.sh -ml

---

Troubleshooting

Browser shows certificate warning

Run the certificate setup manually:

./setup.sh -c

Then restart the server with ./setup.sh -s.

Container won't start

Check the logs to see what went wrong:

./setup.sh -l

Server is unhealthy

./setup.sh -st    # Check health status
./setup.sh -l     # View logs for errors
./setup.sh -r     # Restart the server

Complete reset (start fresh)

This deletes all your data

./setup.sh -t        # Stop the server
docker compose down -v  # Remove all volumes (deletes data)
./setup.sh -s        # Start fresh

---

Next Steps

You're up and running! Now:

• Your First Engagement → — create your first profile and start working
• Managing Profiles → — understand how engagements are organized