Lye

Pre-built tools for AI agents

Lye provides a comprehensive collection of pre-built tools that enable AI agents to interact with the world. From web browsing to file management, image processing to audio transcription, Lye has the tools your agents need.

Overview

Lye is designed to be:

• Easy to use: Import tool groups or individual tools with a single line
• Comprehensive: Covers web, files, images, audio, and more
• Reliable: Battle-tested tools with proper error handling
• Flexible: Use all tools or just the ones you need

Installation

uv add slide-lye

Quick start

from lye import WEB_TOOLS, FILES_TOOLS, IMAGE_TOOLS

# Use with Tyler agents
from tyler import Agent

agent = Agent(
    name="my-agent",
    tools=[*WEB_TOOLS, *FILES_TOOLS]
)

Available tool groups

Web tools

• search - Search the web for information
• fetch - Fetch and parse web pages
• scrape - Extract structured data from websites

File tools

• read_file - Read file contents
• write_file - Write content to files
• list_files - List directory contents
• delete_file - Delete files

Image tools

• describe_image - Get AI descriptions of images
• generate_image - Create images from text
• edit_image - Modify existing images

Audio tools

• transcribe_audio - Convert speech to text
• generate_speech - Convert text to speech

Usage examples

Web research

from lye.web import search, fetch

# Search for information
results = await search("latest AI developments 2024")

# Fetch a specific page
content = await fetch("https://example.com/article")

File operations

from lye.files import read_file, write_file

# Read a file
content = await read_file("data.txt")

# Write processed data
await write_file("results.txt", processed_content)

Multimodal processing

from lye.image import describe_image
from lye.audio import transcribe_audio

# Analyze an image
description = await describe_image("photo.jpg")

# Transcribe audio
transcript = await transcribe_audio("recording.mp3")

Optional dependencies

For advanced file processing capabilities, you can install these system libraries:

macOS

# PDF OCR support for scanned documents
brew install poppler

Ubuntu/Debian

# PDF OCR support for scanned documents
sudo apt-get install poppler-utils

Poppler is only needed for processing scanned PDFs or PDFs without embedded text (OCR).

Next steps

Tool Reference

Detailed documentation for each tool

Adding Tools

Learn to create your own tools