Dashboard

Launch the unified PraisonAI dashboard with Flow, Claw, and Chat UI auto-started

The praisonai dashboard command launches a unified interface that auto-starts Flow Builder, Claw, and Chat UI — one command to open all PraisonAI services.

Quick Start

1. One-command launch

   Launch all services with a single command that auto-opens port 3000 and starts Flow, Claw, and Chat UI:

   praisonai dashboard
   

   This starts the unified dashboard on port 3000 and automatically launches all three services.

2. Standalone dashboard only

   Use when services are already running or you want to manage them manually:

   praisonai dashboard --no-auto-start
   

   This launches only the dashboard interface without starting the backend services.

3. Integrated Agent Dashboard

   Use the new integrated agent host with unified UI pages:

   praisonai dashboard --aiui
   

   This launches the Pattern B in-process host with pages for chat, agents, memory, knowledge, skills, sessions, usage, config, and logs. Requires pip install "praisonai[ui]".

   Set PRAISONAI_HOST_LEGACY=1 for callback-only mode without provider wiring.

---

How It Works

The dashboard follows this process:

Phase	Description
Spawn	Launches Flow (7860), Claw (8082), and UI (8081) services
Readiness	Polls each service for up to 15 seconds until ports accept connections
Dashboard	Starts unified FastAPI interface linking to all services
Cleanup	Gracefully terminates child services on exit

---

Security

The dashboard binds to 127.0.0.1 (localhost only) with no URL-level authentication by default.

Anyone with local machine access can reach the dashboard. Exposing with --host 0.0.0.0 without a reverse proxy or auth layer is not recommended.

For production deployments:

• Use a reverse proxy (nginx, Apache) with authentication
• Consider VPN access for remote management
• Review the Bot Security Guide for additional measures

---

Configuration Options

Option	Type	Default	Description
--port, -p	int	3000	Port to run the unified dashboard on
--host	str	127.0.0.1	Host to bind to (use 0.0.0.0 to expose remotely)
--auto-start / --no-auto-start	bool	True	Auto-start Flow, Claw, and UI services before opening dashboard
--aiui	bool	False	Use integrated agent host (Pattern B) with unified dashboard pages

Ports Used

Service	Default Port	Started by
Unified Dashboard	3000	praisonai dashboard
Flow Builder	7860	Auto-started (praisonai flow --no-open)
Claw Dashboard	8082	Auto-started (praisonai claw)
Chat UI	8081	Auto-started (praisonai ui)

Log Files

Per-service logs are written to:

~/.praisonai/unified/logs/flow.log
~/.praisonai/unified/logs/claw.log
~/.praisonai/unified/logs/ui.log

---

Common Patterns

Expose dashboard on LAN

praisonai dashboard --host 0.0.0.0 --port 9000

Attach to existing services

praisonai dashboard --no-auto-start

Use when Flow, Claw, and UI are already running and you only want the unified interface.

Use integrated agent host

# Install UI dependencies
pip install "praisonai[ui]"

# Launch with integrated dashboard
praisonai dashboard --aiui

This uses Pattern B host integration with unified agent pages. See Host Integration for details.

Troubleshoot failed auto-start

Check service logs if a service fails to start:

# Check individual service logs
tail -f ~/.praisonai/unified/logs/flow.log
tail -f ~/.praisonai/unified/logs/claw.log
tail -f ~/.praisonai/unified/logs/ui.log

---

Which Option Should I Use?

---

Best Practices

Default to unified launch for local development

Use praisonai dashboard as your standard command for local development. It’s designed to feel like launching a single application while giving you access to all PraisonAI interfaces.

Use --no-auto-start in containerized environments

In CI, Docker, or other managed environments where services are started separately, use --no-auto-start to launch only the dashboard interface.

Use --aiui for integrated agent workflows

The --aiui flag launches the new Pattern B host integration with unified agent pages (chat, agents, memory, knowledge, skills, sessions, usage, config, logs). This is now the recommended approach for agent-centric workflows.

Check logs before assuming port conflicts

Before concluding there’s a port conflict, check the per-service logs in ~/.praisonai/unified/logs/ to see the actual error messages from failed services.

---

Related

Claw Dashboard

Full administrative dashboard for managing agents, tools, and channels

Flow Builder

Visual workflow designer with drag-and-drop interface

Chat UI

Clean, distraction-free chat interface

Unified Server

Different feature: HTTP API server (not the desktop launcher)