Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.praison.ai/llms.txt

Use this file to discover all available pages before exploring further.

PraisonAI UI provides a streamlined, single-page chat experience powered by praisonaiui. It removes the complex sidebars and admin panels found in PraisonAI Claw to focus entirely on conversation and agent interaction.

​
Quick Start

1

Install

Install the UI package:
pip install "praisonai[ui]"
2

Launch

Start the clean chat interface:
praisonai ui
The UI will automatically start on http://localhost:8081 and load the default chat application.

​
Features

  • Distraction-Free: No sidebars, no complex configuration pages—just chat.
  • Customizable: Provide your own app.py or agents.yaml file to define custom agents or tasks.
  • Auto-Reload: Supports developer-friendly dynamic reloading.
  • Host Integration: Embed directly in your applications using Pattern B integration.

​
Run in Your Own Host

Integrate PraisonAI UI directly into your Python application: 
from praisonai.integration import build_host_app

# Create host app with UI integration
app = build_host_app(
    title="My Agent App",
    pages=["chat", "agents", "sessions"],
    agents=[{
        "name": "Assistant",
        "instructions": "You are a helpful assistant",
        "llm": "gpt-4o"
    }]
)

# Run with uvicorn, gunicorn, or any ASGI server
# uvicorn main:app --host 0.0.0.0 --port 8000

​
Legacy Mode

praisonai ui now uses configure_host() internally. For the old callback-only behavior: 
export PRAISONAI_HOST_LEGACY=1
praisonai ui
This disables provider wiring and uses only @aiui.reply callbacks. See Host Integration for complete integration patterns.

​
External Agents

The PraisonAI UI automatically detects installed external CLI tools and shows checkboxes for available agents. When working with AgentTeam configurations, enabled external agents add tools that can be merged with YAML-defined role tools.

​
Available Agent Toggles

When external CLIs are installed, the following checkboxes appear automatically:
  • Claude Code: Enable for file editing and advanced coding tasks
  • Gemini CLI: Enable for code analysis and search capabilities
  • Codex CLI: Enable for refactoring and optimization tasks
  • Cursor CLI: Enable for IDE-integrated development workflows

​
AgentTeam Integration

External agent tools merge seamlessly with AgentTeam member configurations: 
# YAML-defined agent with role tools
agent_config = {
    "role": "Senior Developer", 
    "tools": ["file_reader", "web_search"]  # From YAML
}

# External agents add additional tools when toggles are enabled:
# → ["file_reader", "web_search", "claude_code_tool", "gemini_cli_tool"]

​
Session-Based Settings

  • Persistence: Session-based toggles reset on browser refresh
  • Auto-Detection: Only available/installed CLIs show checkboxes
  • Workspace Context: External agents operate within PRAISONAI_WORKSPACE
For comprehensive documentation on external agents across all PraisonAI interfaces, see External Agents in UI.

​
CLI Options

You can customize how the UI starts using flags: 
praisonai ui                        # Chat on port 8081
praisonai ui --port 9000            # Run on custom port 9000
praisonai ui --host 127.0.0.1      # Bind to localhost only
praisonai ui --app my-chat.py       # Load a custom Python config
praisonai ui --reload               # Enable auto-reload on file changes

​
Custom Chat Applications

By default, the praisonai ui command provisions a default chat application located at ~/.praisonai/ui/app.py. You can define your own custom chat experience using a standard Python script and pointing the praisonai ui command to it: 
praisonai ui --app custom-agent.py

PraisonAI Claw

Need a full administrative dashboard? Use PraisonAI Claw to manage channels, knowledge, and multi-agent memory.

PraisonAI Chat

Want to dynamically chat with 100+ LLMs with vision capabilities? Explore PraisonAI Chat.