Deployment Modes

The BugTraceAI-Launcher supports three deployment modes, each targeting different use cases. This page describes each mode in detail, including what services are deployed, how they communicate, and when to use each one.

Mode Comparison

Feature	Full Platform	Standalone WEB	Standalone CLI
WEB Dashboard	Yes	Yes	No
CLI Scanner	Yes	No	Yes
MCP for AI Assistants	Yes	No	Yes
Real-time Monitoring	Yes	No	N/A
Security Toolkit (20+ tools)	Yes	Yes	No
Active Scanning	Yes	No	Yes
REST API	Yes	No	Yes
PostgreSQL	Yes	Yes	No
CI/CD Integration	Yes	No	Yes
Auto-connected	Yes	N/A	N/A

The MCP server (bugtrace_mcp, port 8001) ships with the CLI container, so it is available in both Full Platform and Standalone CLI — there is no separate “CLI + AI Assistant” deployment mode.

Full Platform

Command: ./launcher.sh then select “Full Platform”

Description

Deploys both WEB and CLI, automatically configures the connection between them, and starts all supporting services. This is the recommended mode for most users.

Services

Service	Port	Container	Description
CLI FastAPI	8000	`btai-cli`	Scanning engine and REST API
WEB Frontend	6869	`btai-web`	React dashboard (Nginx)
WEB Backend	3001	`btai-backend`	Express API
PostgreSQL	5432 (internal)	`btai-postgres`	WEB database

Auto-Configuration

The Launcher automatically sets:

VITE_CLI_API_URL=/cli-api on the WEB frontend (proxied through Nginx)
CORS headers on the CLI to allow requests from the WEB origin
Docker networking for inter-container communication
PostgreSQL credentials and database initialization

Data Flow

User --> WEB Dashboard (:6869)
             |
             +-- REST API --> CLI (:8000) --> SQLite
             |
             +-- WebSocket --> CLI (:8000) --> Event Bus
             |
             +-- HTTP --> WEB Backend (:3001) --> PostgreSQL

In Docker deployments, the WEB frontend’s Nginx reverse proxy routes /cli-api/ to the CLI FastAPI server. This eliminates CORS issues and simplifies networking.

Best For

Complete pentesting platform
Bug bounty workflows
Security team deployments
Users who want everything in one setup

Standalone WEB

Command: ./launcher.sh then select “Standalone WEB”

Description

Deploys only the WEB dashboard with its backend and database. Provides access to the 20+ AI-powered security toolkit tools without active scanning capabilities.

Services

Service	Port	Container	Description
WEB Frontend	6869	`btai-web`	React dashboard (Nginx)
WEB Backend	3001	`btai-backend`	Express API
PostgreSQL	5432 (internal)	`btai-postgres`	WEB database

What You Get

All 20+ security toolkit tools (JWT Analyzer, Payload Forge, Code Analyzer, etc.)
Chat-based AI interaction for security analysis
Persistent storage of analyses and conversations
No CLI scanner, no active scanning, no vulnerability exploitation

What You Do NOT Get

No active scanning capability
No real-time scan monitoring
No vulnerability exploitation or validation
No report generation from scans

Best For

AI-assisted security analysis only
Code review and vulnerability research
Learning security concepts
Environments where active scanning is not permitted

Standalone CLI

Command: ./launcher.sh then select “Standalone CLI”

Description

Deploys only the CLI scanning engine as a headless API server. Provides full scanning capability via REST API and WebSocket without any graphical interface.

Services

Service	Port	Container	Description
CLI FastAPI	8000	`btai-cli`	Scanning engine and REST API

What You Get

Complete autonomous scanning engine
REST API for scan management
WebSocket for real-time event streaming
All specialist agents and Go fuzzers
Browser-based validation (Playwright)
Report generation (HTML, JSON, Markdown)
SQLite persistence

What You Do NOT Get

No graphical dashboard
No security toolkit tools
No chat-based AI interaction
No PostgreSQL

API Access

# Start a scan
curl -X POST http://localhost:8000/api/scans \
  -H "Content-Type: application/json" \
  -d '{"target_url": "https://example.com"}'

# Check status
curl http://localhost:8000/api/scans/{id}/status

# Download report
curl http://localhost:8000/api/scans/{id}/report/json > report.json

Best For

CI/CD pipeline integration
Scripted and automated scanning
Headless server environments
API-only access
Integration with custom dashboards or tools

AI Assistant Control (built into the CLI)

Command: ./launcher.sh then select Standalone CLI (or Full Platform) — no separate selection is needed.

Description

The CLI container always starts an MCP (Model Context Protocol) server alongside the scanning engine, so every CLI deployment can be controlled from an AI assistant like OpenClaw, Claude Code, Cursor, or any MCP-compatible client. This is not a distinct deployment mode — it is a capability included with the CLI.

Services

Service	Port	Container	Description
CLI FastAPI	8000	`bugtrace_api`	Scanning engine and REST API
MCP Server	8001	`bugtrace_mcp`	SSE transport for AI assistants

What You Get

Complete autonomous scanning engine
REST API for scan management
MCP tools accessible from AI assistants (start scans, check status, get findings, export reports)
All specialist agents and Go fuzzers
Browser-based validation (Playwright)
Report generation (HTML, JSON, Markdown)
SQLite persistence

What You Do NOT Get

No graphical dashboard
No security toolkit tools
No PostgreSQL

MCP Configuration

After deployment, add BugTraceAI to your AI assistant’s MCP configuration:

{
  "mcpServers": {
    "bugtraceai": {
      "baseUrl": "http://localhost:8001/sse",
      "description": "BugTraceAI Security Scanner"
    }
  }
}

Best For

AI-driven security testing via Telegram (OpenClaw), terminal (Claude Code), or IDE (Cursor)
Users who prefer chat-based interfaces over dashboards
Automated scanning controlled by AI agents
Integration with AI workflows and assistants

See AI Assistant Integration for full setup instructions and available MCP tools.

Choosing a Mode

Do you need a graphical dashboard?
  |
  +-- No --> Do you need active scanning?
  |            |
  |            +-- Yes --> Standalone CLI   (includes the MCP server for AI-assistant control)
  |            +-- No  --> (You may not need BugTraceAI)
  |
  +-- Yes --> Do you need active scanning?
               |
               +-- Yes --> Full Platform
               +-- No  --> Standalone WEB

Switching Modes

You can switch between modes by stopping the current deployment and running the Launcher wizard again:

./launcher.sh stop
./launcher.sh
# Select a different mode

Data is preserved between mode switches:

PostgreSQL data persists in Docker volumes
SQLite data persists in the CLI container volume
Configuration is regenerated for the new mode

Manual Deployment

If you prefer to deploy without the Launcher, each component can be deployed independently:

Manual CLI Deployment

cd BugTraceAI-CLI
docker build -t bugtrace-cli .
docker run -d -p 8000:8000 --name btai-cli bugtrace-cli

Manual WEB Deployment

cd BugTraceAI-WEB
docker build -t bugtrace-web .
docker run -d -p 6869:6869 --name btai-web bugtrace-web

Manual Full Stack

See the Docker Compose configuration examples in the Architecture page.

Production Considerations

For production deployments:

Concern	Recommendation
SSL/TLS	Place Nginx reverse proxy with Let’s Encrypt in front
Firewall	Allow only ports 6869 and 8000 externally
Backups	Schedule regular PostgreSQL and SQLite backups
Updates	Use `./launcher.sh update` for rolling updates
Monitoring	Use `./launcher.sh status` and `./launcher.sh logs`
Resources	8 GB RAM recommended for Full Platform mode

Parent: BugTraceAI-Launcher