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.

Streaming CLI

Control streaming behavior when running agents from the command line.

Overview

By default, agent responses stream to the terminal in real-time. You can control this behavior with CLI flags.

Commands

Enable Streaming (Default)

praisonai-ts chat "Tell me a story" --stream

Disable Streaming

praisonai-ts chat "Tell me a story" --no-stream

Options

OptionDescriptionDefault
--streamEnable streaming outputtrue
--no-streamDisable streaming, wait for complete responsefalse
--verbose, -vShow streaming output in terminaltrue
--quiet, -qSuppress streaming outputfalse

Examples

Interactive Streaming

# Stream response as it's generated
praisonai-ts chat "Explain quantum computing in detail" --stream

Batch Processing

# Get complete response (better for scripting)
praisonai-ts chat "Summarize this text" --no-stream --json

Silent Mode

# No streaming output, just final result
praisonai-ts chat "Hello" --quiet

Verbose Streaming

# Full streaming with debug info
praisonai-ts chat "Hello" --stream --verbose

Scripting Examples

Capture Streamed Output

#!/bin/bash

# Stream to file
praisonai-ts chat "Write a long story" --stream > story.txt

# Stream and display
praisonai-ts chat "Explain AI" --stream | tee output.txt

JSON Output (No Streaming)

# For API-like usage, disable streaming
RESULT=$(praisonai-ts chat "Hello" --no-stream --json)
echo $RESULT | jq '.data.response'

Pipeline Usage

# Pipe between commands (streaming disabled automatically)
echo "Summarize this" | praisonai-ts chat --stdin --no-stream

Environment Variables

# Default streaming behavior
export PRAISON_STREAM=true

# Default verbose behavior
export PRAISON_VERBOSE=true

Use Cases

Interactive Chat

# Best experience for interactive use
praisonai-ts chat "Let's have a conversation" --stream --verbose

Automation Scripts

# Best for scripts and automation
praisonai-ts chat "Process this data" --no-stream --json

Long-Form Generation

# Stream long content to see progress
praisonai-ts run \
  --instructions "You are a writer" \
  --task "Write a 2000-word essay" \
  --stream

See Also