Skip to main content
Slash commands and CLI flags control models, sessions, and layout inside the TUI. 
from praisonaiagents import Agent

agent = Agent(name="tui-assistant", instructions="Use TUI slash commands effectively.")
agent.start("Which /model should I pick for fast replies?")
The user types slash commands such as /help and /model while the agent streams responses in the TUI.

​
TUI CLI Commands

Complete reference for all TUI-related CLI commands.

​
Quick Start

1

Launch the TUI

praisonai tui launch --model gpt-4
2

Run a headless simulation

praisonai tui simulate test_script.yaml --assert

​
praisonai tui

Main TUI command group for interactive terminal interface.

​
launch

Launch the interactive TUI. 
praisonai tui launch [OPTIONS]
Options:
OptionShortDescription
--workspace-wWorkspace directory
--session-sResume session ID
--model-mDefault model
--debug-dEnable debug overlays
--log-jsonlWrite events to JSONL file
--profileEnable performance profiling
Examples: 
# Basic launch
praisonai tui launch

# Launch with specific model
praisonai tui launch --model gpt-4

# Resume a session
praisonai tui launch --session abc123

# Launch with debug logging
praisonai tui launch --debug --log-jsonl ./events.jsonl

​
simulate

Run a headless TUI simulation script for testing. 
praisonai tui simulate SCRIPT [OPTIONS]
Arguments:
ArgumentDescription
SCRIPTPath to simulation script (YAML/JSON)
Options:
OptionDescription
--mock/--real-llmUse mock provider (default) or real LLM
--pretty/--jsonlOutput format
--assertValidate expected outcomes
--timeoutMax execution time (default: 60s)
Examples: 
# Run simulation with mock provider
praisonai tui simulate test_script.yaml

# Run with real LLM (requires PRAISONAI_REAL_LLM=1)
PRAISONAI_REAL_LLM=1 praisonai tui simulate test_script.yaml --real-llm

# Run with assertions
praisonai tui simulate test_script.yaml --assert

​
snapshot

Print a TUI-like snapshot of current state. 
praisonai tui snapshot [OPTIONS]
Options:
OptionShortDescription
--session-sFilter by session ID
--run-rFilter by run ID
--json-jOutput as JSON
Examples: 
# Get current snapshot
praisonai tui snapshot

# Get snapshot for specific session
praisonai tui snapshot --session abc123

# Get JSON output
praisonai tui snapshot --json

​
trace

Replay events from persistence like a timeline. 
praisonai tui trace ID [OPTIONS]
Arguments:
ArgumentDescription
IDSession or run ID to trace
Options:
OptionShortDescription
--follow-fFollow new events
--limit-nMax events to show (default: 50)
Examples: 
# Trace a session
praisonai tui trace abc123

# Follow events in real-time
praisonai tui trace abc123 --follow

# Limit to last 10 events
praisonai tui trace abc123 --limit 10

​
praisonai queue

Queue management commands.

​
ls

List queued runs. 
praisonai queue ls [OPTIONS]
Options:
OptionShortDescription
--state-sFilter by state (queued, running, succeeded, failed, cancelled)
--sessionFilter by session ID
--limit-nMaximum results (default: 20)
--json-jOutput as JSON
Examples: 
# List all runs
praisonai queue ls

# List only running
praisonai queue ls --state running

# List with JSON output
praisonai queue ls --json

​
cancel

Cancel a queued or running run. 
praisonai queue cancel RUN_ID
Arguments:
ArgumentDescription
RUN_IDRun ID to cancel (partial match supported)
Examples: 
# Cancel by full ID
praisonai queue cancel abc12345

# Cancel by partial ID
praisonai queue cancel abc1

​
retry

Retry a failed run. 
praisonai queue retry RUN_ID
Arguments:
ArgumentDescription
RUN_IDRun ID to retry
Examples: 
praisonai queue retry abc12345

​
clear

Clear all queued runs. 
praisonai queue clear [OPTIONS]
Options:
OptionShortDescription
--force-fSkip confirmation
Examples: 
# Clear with confirmation
praisonai queue clear

# Force clear
praisonai queue clear --force

​
stats

Show queue statistics. 
praisonai queue stats [OPTIONS]
Options:
OptionDescription
--sessionFilter by session ID
Examples: 
# Show all stats
praisonai queue stats

# Show stats for session
praisonai queue stats --session abc123

​
Environment Variables

VariableDescription
PRAISONAI_REAL_LLMSet to 1 to enable real LLM in simulations
PRAISONAI_TUI_DEBUGSet to 1 to enable debug mode
PRAISONAI_TUI_JSONLPath for JSONL event logging

​
Best Practices

Run praisonai simulate with mock providers to avoid live API spend in pipelines.
Set PRAISONAI_TUI_JSONL when you need structured event replay after failures.
PRAISONAI_TUI_DEBUG=1 is for local troubleshooting only — it adds verbose noise.
Share a cheatsheet of /context, queue, and simulate commands so operators behave consistently.

TUI Overview

Interactive terminal interface

TUI Simulation

Headless testing and CI