AI Agent Tools Reference Guide

A comprehensive guide to tools available for LangChain/LangGraph agents, organized by deployment type and cost structure.

1. Free Tools That Run Locally

Python REPL Tool

Description: Allows agents to execute Python code dynamically in a sandboxed environment. Essential for mathematical calculations, data analysis, and computational tasks.

Documentation: https://python.langchain.com/docs/integrations/tools/python

Installation:


pip install langchain-experimental

LangChain Integration: ✅ Pre-built tool available (PythonREPLTool)

File Management Tools (Read/Write/List)

Description: Suite of tools for file operations including reading file contents, writing new files, listing directories, and moving files.

Documentation: https://python.langchain.com/docs/integrations/tools/filesystem

Installation:


pip install langchain-community

LangChain Integration: ✅ Pre-built tools available (ReadFileTool, WriteFileTool, ListDirectoryTool, MoveFileTool)

DuckDuckGo Search

Description: Privacy-focused web search with no API key required. Returns raw search results that agents can parse.

Documentation: https://python.langchain.com/docs/integrations/tools/ddg

Installation:


pip install duckduckgo-search langchain-community

LangChain Integration: ✅ Pre-built tool available (DuckDuckGoSearchRun)

Wikipedia Tool

Description: Direct access to Wikipedia content for encyclopedic knowledge and reliable background information.

Documentation: https://python.langchain.com/docs/integrations/tools/wikipedia

Installation:


pip install wikipedia langchain-community

LangChain Integration: ✅ Pre-built tool available (WikipediaQueryRun)

ArXiv Tool

Description: Search and retrieve scientific papers from ArXiv for research-focused applications.

Documentation: https://python.langchain.com/docs/integrations/tools/arxiv

Installation:


pip install arxiv langchain-community

LangChain Integration: ✅ Pre-built tool available (ArxivQueryRun)

Shell Tool

Description: Execute bash commands in the system shell for file system operations, process management, and DevOps tasks.

Documentation: https://python.langchain.com/docs/integrations/tools/bash

Installation:


pip install langchain-community

LangChain Integration: ✅ Pre-built tool available (ShellTool)

⚠️ Security Warning: Requires careful sandboxing in production environments.

Requests Tool

Description: Make HTTP requests to external APIs and web services.

Documentation: https://python.langchain.com/docs/integrations/tools/requests

Installation:


pip install requests langchain-community

LangChain Integration: ✅ Pre-built toolkit available (RequestsToolkit)

SQL Database Tool

Description: Query SQL databases using natural language. Agent can inspect schema, construct queries, and interpret results.

Documentation: https://python.langchain.com/docs/integrations/tools/sql_database

Installation:


pip install langchain-community sqlalchemy
# Plus database-specific driver (e.g., psycopg2, pymysql, sqlite3)

LangChain Integration: ✅ Pre-built toolkit available (SQLDatabaseToolkit)

Pandas DataFrame Tool

Description: Manipulate pandas DataFrames using natural language for data analysis, filtering, aggregation, and transformation.

Documentation: https://python.langchain.com/docs/integrations/tools/pandas

Installation:


xxxxxxxxxx
pip install pandas langchain-experimental

LangChain Integration: ✅ Pre-built tool available (PythonAstREPLTool with DataFrame support)

YouTube Transcript API

Description: Fetch transcripts/captions from YouTube videos. No API key required. Useful for video content analysis.

Documentation: https://github.com/jdepoix/youtube-transcript-api

Installation:


xxxxxxxxxx
pip install youtube-transcript-api

LangChain Integration: ❌ Custom tool definition required

Example Tool Definition:


x
from langchain.tools import tool
from youtube_transcript_api import YouTubeTranscriptApi

@tool
def get_youtube_transcript(video_id: str) -> str:
    """Get transcript from a YouTube video given its video ID."""
    transcript = YouTubeTranscriptApi.get_transcript(video_id)
    return " ".join([entry['text'] for entry in transcript])

Pillow (PIL)

Description: Python's standard image processing library for resize, crop, rotate, format conversion, and basic image manipulation.

Documentation: https://pillow.readthedocs.io/

Installation:


xxxxxxxxxx
pip install Pillow

LangChain Integration: ❌ Custom tool definition required

Example Tool Definition:


xxxxxxxxxx
from langchain.tools import tool
from PIL import Image

@tool
def resize_image(image_path: str, width: int, height: int, output_path: str) -> str:
    """Resize an image to specified dimensions."""
    img = Image.open(image_path)
    img_resized = img.resize((width, height))
    img_resized.save(output_path)
    return f"Image resized and saved to {output_path}"

OpenCV (cv2)

Description: Industry-standard computer vision library for advanced image/video analysis, object detection, and feature extraction.

Documentation: https://docs.opencv.org/

Installation:


xxxxxxxxxx
pip install opencv-python

LangChain Integration: ❌ Custom tool definition required

imageio

Description: Simple library for reading/writing image and video data, particularly good for video frame extraction.

Documentation: https://imageio.readthedocs.io/

Installation:


xxxxxxxxxx
pip install imageio imageio-ffmpeg

LangChain Integration: ❌ Custom tool definition required

Tesseract OCR

Description: Open-source OCR engine for extracting text from images. Supports 100+ languages.

Documentation: https://github.com/tesseract-ocr/tesseract

Installation:


xxxxxxxxxx
# Install Tesseract binary first (OS-specific)
# Ubuntu: sudo apt-get install tesseract-ocr
# Mac: brew install tesseract
# Windows: Download installer from GitHub

pip install pytesseract

LangChain Integration: ❌ Custom tool definition required

Example Tool Definition:


xxxxxxxxxx
from langchain.tools import tool
import pytesseract
from PIL import Image

@tool
def extract_text_from_image(image_path: str) -> str:
    """Extract text from an image using OCR."""
    img = Image.open(image_path)
    text = pytesseract.image_to_string(img)
    return text

EasyOCR

Description: Deep learning-based OCR with better accuracy than Tesseract on complex images. Supports 80+ languages.

Documentation: https://github.com/JaidedAI/EasyOCR

Installation:


xxxxxxxxxx
pip install easyocr

LangChain Integration: ❌ Custom tool definition required

PyPDF / PDFPlumber

Description: Extract text and tables from PDF files. PyPDF for basic extraction, PDFPlumber for layout-aware parsing.

Documentation:

PyPDF: https://pypdf2.readthedocs.io/
PDFPlumber: https://github.com/jsvine/pdfplumber

Installation:


xxxxxxxxxx
pip install pypdf2 pdfplumber

LangChain Integration: ✅ Pre-built loader available (PyPDFLoader), but not as a tool. Custom tool definition recommended for agent use.

Beautiful Soup

Description: Web scraping library for parsing HTML and extracting data from static web pages.

Documentation: https://www.crummy.com/software/BeautifulSoup/bs4/doc/

Installation:


xxxxxxxxxx
pip install beautifulsoup4 requests

LangChain Integration: ❌ Custom tool definition required

Example Tool Definition:


xxxxxxxxxx
from langchain.tools import tool
from bs4 import BeautifulSoup
import requests

@tool
def scrape_webpage(url: str) -> str:
    """Scrape and extract text content from a webpage."""
    response = requests.get(url)
    soup = BeautifulSoup(response.content, 'html.parser')
    return soup.get_text()

Playwright

Description: Modern browser automation for JavaScript-heavy websites. Best headless browser option for LangChain agents.

Documentation: https://playwright.dev/python/

Installation:


xxxxxxxxxx
pip install playwright langchain-community
playwright install  # Downloads browser binaries

LangChain Integration: ✅ Pre-built toolkit available (PlaywrightBrowserToolkit)

Selenium

Description: Mature browser automation tool with wide browser support. Slower than Playwright but larger ecosystem.

Documentation: https://selenium-python.readthedocs.io/

Installation:


xxxxxxxxxx
pip install selenium webdriver-manager

LangChain Integration: ✅ Pre-built toolkit available (SeleniumToolkit in langchain-community)

Requests-HTML

Description: Lightweight library combining Requests with Chromium rendering for basic JavaScript rendering.

Documentation: https://requests.readthedocs.io/projects/requests-html/

Installation:


xxxxxxxxxx
pip install requests-html

LangChain Integration: ❌ Custom tool definition required

MoviePy

Description: Python library for programmatic video editing including cutting, concatenating, and adding effects.

Documentation: https://zulko.github.io/moviepy/

Installation:


xxxxxxxxxx
pip install moviepy

LangChain Integration: ❌ Custom tool definition required

FFmpeg (via subprocess)

Description: Command-line tool for video/audio processing. Handles virtually any codec or format.

Documentation: https://ffmpeg.org/documentation.html

Installation:


xxxxxxxxxx
# Install FFmpeg binary (OS-specific)
# Ubuntu: sudo apt-get install ffmpeg
# Mac: brew install ffmpeg
# Windows: Download from ffmpeg.org

# No Python package needed - call via subprocess

LangChain Integration: ❌ Custom tool definition required

2. Free OAuth Tools

These tools require user authentication but have no per-use costs beyond OAuth setup.

Gmail Tool

Description: Read emails, send messages, search inbox, and manage labels through Gmail API.

Documentation: https://python.langchain.com/docs/integrations/tools/gmail

Installation:


xxxxxxxxxx
pip install google-auth google-auth-oauthlib google-auth-httplib2 google-api-python-client langchain-google-community

LangChain Integration: ✅ Pre-built toolkit available (GmailToolkit)

Setup Required: Google Cloud project, OAuth 2.0 credentials, user consent

Google Calendar Tool

Description: Create events, check availability, update meetings, and manage calendars.

Documentation: https://python.langchain.com/docs/integrations/tools/google_calendar

Installation:


xxxxxxxxxx
pip install google-auth google-auth-oauthlib google-auth-httplib2 google-api-python-client langchain-google-community

LangChain Integration: ✅ Pre-built toolkit available (GoogleCalendarToolkit)

Setup Required: Google Cloud project, OAuth 2.0 credentials, user consent

Google Sheets API

Description: Read/write spreadsheets, create charts, and collaborate in familiar tools.

Documentation: https://developers.google.com/sheets/api

Installation:


xxxxxxxxxx
pip install google-auth google-auth-oauthlib google-auth-httplib2 google-api-python-client gspread

LangChain Integration: ❌ Custom tool definition required (no official LangChain tool)

GitHub API

Description: Create issues, make pull requests, analyze code, and manage repositories.

Documentation: https://python.langchain.com/docs/integrations/tools/github

Installation:


xxxxxxxxxx
pip install PyGithub langchain-community

LangChain Integration: ✅ Pre-built toolkit available (GitHubToolkit)

Setup Required: GitHub personal access token or OAuth app

Slack API

Description: Send messages, read channels, create threads, and respond to mentions.

Documentation: https://python.langchain.com/docs/integrations/tools/slack

Installation:


xxxxxxxxxx
pip install slack-sdk langchain-community

LangChain Integration: ✅ Pre-built toolkit available (SlackToolkit)

Setup Required: Slack app creation, bot token

Discord API

Description: Similar capabilities to Slack but for Discord servers. Send messages, manage channels, respond to events.

Documentation: https://discordpy.readthedocs.io/

Installation:


xxxxxxxxxx
pip install discord.py

LangChain Integration: ❌ Custom tool definition required

Notion API

Description: Read/write pages, databases, and blocks for knowledge management and documentation.

Documentation: https://developers.notion.com/

Installation:


xxxxxxxxxx
pip install notion-client

LangChain Integration: ❌ Custom tool definition required (loaders available but not tools)

Jira API

Description: Create tickets, update issues, track project status for project management automation.

Documentation: https://python.langchain.com/docs/integrations/tools/jira

Installation:


xxxxxxxxxx
pip install jira langchain-community

LangChain Integration: ✅ Pre-built toolkit available (JiraToolkit)

Setup Required: Jira instance URL, API token or OAuth

3. Limited-Tier APIs

These services offer free tiers with usage limits. Monitor usage carefully in classroom settings.

Tavily Search

Description: AI-optimized web search returning clean, structured results. Best modern search option for agents.

Documentation: https://python.langchain.com/docs/integrations/tools/tavily_search

Installation:


xxxxxxxxxx
pip install tavily-python langchain-community

LangChain Integration: ✅ Pre-built tool available (TavilySearchResults)

Free Tier: 1,000 searches/month
API Key Required: Yes (free at tavily.com)

Serper (Google Search)

Description: Google search results via API with structured output including snippets and knowledge graphs.

Documentation: https://python.langchain.com/docs/integrations/tools/google_serper

Installation:


xxxxxxxxxx
pip install google-search-results langchain-community

LangChain Integration: ✅ Pre-built tool available (GoogleSerperAPIWrapper)

Free Tier: 2,500 queries/month
API Key Required: Yes (free at serper.dev)

SerpAPI

Description: Multi-engine search wrapper (Google, Bing, DuckDuckGo) with specialized searches.

Documentation: https://python.langchain.com/docs/integrations/tools/serpapi

Installation:


xxxxxxxxxx
pip install google-search-results langchain-community

LangChain Integration: ✅ Pre-built tool available (SerpAPIWrapper)

Free Tier: 100 searches/month
API Key Required: Yes (free at serpapi.com)

OpenWeatherMap

Description: Weather data including current conditions, forecasts, and historical data.

Documentation: https://openweathermap.org/api

Installation:


xxxxxxxxxx
pip install pyowm

LangChain Integration: ✅ Pre-built tool available (OpenWeatherMapQueryRun)

Free Tier: 1,000 calls/day
API Key Required: Yes (free at openweathermap.org)

Alpha Vantage (Stock Market)

Description: Stock prices, forex rates, cryptocurrency data, and technical indicators.

Documentation: https://www.alphavantage.co/documentation/

Installation:


xxxxxxxxxx
pip install alpha-vantage

LangChain Integration: ❌ Custom tool definition required

Free Tier: 25 API calls/day
API Key Required: Yes (free at alphavantage.co)

NewsAPI

Description: Aggregated news articles from 80,000+ sources worldwide with metadata.

Documentation: https://newsapi.org/docs

Installation:


xxxxxxxxxx
pip install newsapi-python

LangChain Integration: ❌ Custom tool definition required

Free Tier: 100 requests/day (developer plan)
API Key Required: Yes (free at newsapi.org)

Wolfram Alpha API

Description: Computational knowledge engine for math, science, and factual queries.

Documentation: https://products.wolframalpha.com/api/documentation

Installation:


xxxxxxxxxx
pip install wolframalpha langchain-community

LangChain Integration: ✅ Pre-built tool available (WolframAlphaQueryRun)

Free Tier: 2,000 queries/month
API Key Required: Yes (free at developer.wolframalpha.com)

Metaphor Search (Exa)

Description: AI-powered semantic search for finding high-quality, contextually relevant content.

Documentation: https://docs.exa.ai/

Installation:


xxxxxxxxxx
pip install exa-py

LangChain Integration: ✅ Pre-built tool available (ExaSearchResults in langchain-exa)

Free Tier: 1,000 searches/month
API Key Required: Yes (free at exa.ai)

Zapier Natural Language Actions

Description: Trigger 5,000+ app integrations through natural language (Slack, Gmail, Sheets, Notion, etc.).

Documentation: https://python.langchain.com/docs/integrations/tools/zapier

Installation:


xxxxxxxxxx
pip install langchain-community

LangChain Integration: ✅ Pre-built toolkit available (ZapierToolkit)

Free Tier: 100 tasks/month
API Key Required: Yes (requires Zapier account)

Unstructured.io

Description: Extract clean text and structured data from PDFs, Word docs, PowerPoints, images, and more.

Documentation: https://unstructured-io.github.io/unstructured/

Installation:


xxxxxxxxxx
pip install unstructured
# Or use API:
pip install unstructured-client

LangChain Integration: ✅ Pre-built loader available (UnstructuredFileLoader)

Free Tier: 1,000 pages/month (API)
API Key Required: Yes for API (free at unstructured.io)

Apify

Description: Web scraping platform with pre-built scrapers for social media, e-commerce, etc.

Documentation: https://docs.apify.com/

Installation:


xxxxxxxxxx
pip install apify-client langchain-community

LangChain Integration: ✅ Pre-built wrapper available (ApifyWrapper)

Free Tier: $5/month credit
API Key Required: Yes (free at apify.com)

Hugging Face Inference API

Description: Access thousands of ML models for classification, NER, sentiment, image processing.

Documentation: https://python.langchain.com/docs/integrations/tools/huggingface_tools

Installation:


xxxxxxxxxx
pip install huggingface-hub langchain-community

LangChain Integration: ✅ Pre-built tools available (HuggingFaceHub, load_huggingface_tool)

Free Tier: Rate-limited free access
API Key Required: Yes (free at huggingface.co)

Airtable API

Description: Database/spreadsheet hybrid for structured data with relationships.

Documentation: https://airtable.com/developers/web/api/introduction

Installation:


xxxxxxxxxx
pip install pyairtable

LangChain Integration: ❌ Custom tool definition required

Free Tier: 1,200 records per base
API Key Required: Yes (free Airtable account)

Twilio (SMS/Voice)

Description: Send SMS, make phone calls, send WhatsApp messages.

Documentation: https://www.twilio.com/docs/usage/api

Installation:


xxxxxxxxxx
pip install twilio

LangChain Integration: ❌ Custom tool definition required

Free Tier: $15 trial credit
API Key Required: Yes (trial at twilio.com)

SendGrid (Email)

Description: Send transactional emails and newsletters programmatically.

Documentation: https://docs.sendgrid.com/

Installation:


xxxxxxxxxx
pip install sendgrid

LangChain Integration: ❌ Custom tool definition required

Free Tier: 100 emails/day
API Key Required: Yes (free at sendgrid.com)

Linear API

Description: Modern project management with cleaner API than Jira for task and sprint management.

Documentation: https://developers.linear.app/

Installation:


xxxxxxxxxx
pip install linear-sdk

LangChain Integration: ❌ Custom tool definition required

Free Tier: Unlimited API access (with Linear account)
API Key Required: Yes (free Linear account)

OpenCage Geocoding

Description: Convert addresses to coordinates and vice versa.

Documentation: https://opencagedata.com/api

Installation:


xxxxxxxxxx
pip install opencage

LangChain Integration: ❌ Custom tool definition required

Free Tier: 2,500 requests/day
API Key Required: Yes (free at opencagedata.com)

Nominatim (OpenStreetMap Geocoding)

Description: Completely free geocoding service from OpenStreetMap.

Documentation: https://nominatim.org/release-docs/develop/api/Overview/

Installation:


xxxxxxxxxx
pip install geopy

LangChain Integration: ❌ Custom tool definition required

Free Tier: Unlimited (with fair use policy)
API Key Required: No

Google Cloud Vision API

Description: Comprehensive image analysis: object detection, OCR, face detection, landmark recognition.

Documentation: https://cloud.google.com/vision/docs

Installation:


xxxxxxxxxx
pip install google-cloud-vision

LangChain Integration: ❌ Custom tool definition required

Free Tier: 1,000 units/month for most features
API Key Required: Yes (Google Cloud account)

AWS Rekognition

Description: Image and video analysis with strong video capabilities and celebrity recognition.

Documentation: https://docs.aws.amazon.com/rekognition/

Installation:


xxxxxxxxxx
pip install boto3

LangChain Integration: ❌ Custom tool definition required

Free Tier: 5,000 images/month (first 12 months)
API Key Required: Yes (AWS account)

Azure Computer Vision

Description: Microsoft's vision API with strong document understanding and form processing.

Documentation: https://learn.microsoft.com/en-us/azure/ai-services/computer-vision/

Installation:


xxxxxxxxxx
pip install azure-cognitiveservices-vision-computervision

LangChain Integration: ❌ Custom tool definition required

Free Tier: 5,000 transactions/month
API Key Required: Yes (Azure account)

Roboflow

Description: Platform for training custom object detection models with your own data.

Documentation: https://docs.roboflow.com/

Installation:


xxxxxxxxxx
pip install roboflow

LangChain Integration: ❌ Custom tool definition required

Free Tier: Available with limits
API Key Required: Yes (free at roboflow.com)

Cloudinary

Description: Media management with image/video transformation, optimization, and content-aware resizing.

Documentation: https://cloudinary.com/documentation

Installation:


xxxxxxxxxx
pip install cloudinary

LangChain Integration: ❌ Custom tool definition required

Free Tier: 25 GB storage, 25 GB bandwidth/month
API Key Required: Yes (free at cloudinary.com)

DALL-E (OpenAI)

Description: Generate images from text descriptions using OpenAI's image generation model.

Documentation: https://platform.openai.com/docs/guides/images

Installation:


xxxxxxxxxx
pip install openai

LangChain Integration: ✅ Pre-built tool available (DallEAPIWrapper in langchain-community)

Free Tier: Free trial credits
API Key Required: Yes (OpenAI account with credits)

Stability AI (DreamStudio)

Description: Image generation using Stable Diffusion models.

Documentation: https://platform.stability.ai/docs

Installation:


xxxxxxxxxx
pip install stability-sdk

LangChain Integration: ❌ Custom tool definition required

Free Tier: $5 free credit
API Key Required: Yes (free at dreamstudio.ai)

ElevenLabs (Text-to-Speech)

Description: High-quality voice synthesis for natural-sounding speech from text.

Documentation: https://elevenlabs.io/docs

Installation:


xxxxxxxxxx
pip install elevenlabs

LangChain Integration: ❌ Custom tool definition required

Free Tier: 10,000 characters/month
API Key Required: Yes (free at elevenlabs.io)

AssemblyAI (Speech-to-Text)

Description: Audio transcription with speaker identification and content analysis.

Documentation: https://www.assemblyai.com/docs

Installation:


xxxxxxxxxx
pip install assemblyai

LangChain Integration: ❌ Custom tool definition required

Free Tier: Available for testing
API Key Required: Yes (free at assemblyai.com)

Quick Reference Table

Tool Category	Recommended Starter Tools	Key Advantage
Web Search	Tavily, DuckDuckGo	Tavily for quality, DDG for zero cost
Document Processing	PyPDF, Unstructured.io	Local first, API for complex formats
Web Scraping	Playwright, Beautiful Soup	Playwright for JS sites, BS4 for static
Image Processing	Pillow, OpenCV	Start simple (Pillow), graduate to advanced (OpenCV)
OCR	Tesseract, EasyOCR	Tesseract for basic, EasyOCR for accuracy
Communication	Gmail, Slack	Students likely have accounts
Code Execution	Python REPL	Essential for computational agents
Data Analysis	Pandas DataFrame, SQL Database	Core tools for data-focused agents

Classroom Deployment Recommendations

Week 1-2: Local Tools Only

Python REPL, File Tools, Pandas DataFrame
Cost: $0
Goal: Understand tool interface fundamentals

Week 3-4: Free OAuth Tools

Gmail, Google Sheets, GitHub
Cost: $0
Goal: Learn authentication patterns

Week 5-6: Web Interaction

DuckDuckGo Search, Playwright, Beautiful Soup
Cost: $0
Goal: Build web-aware agents

Week 7+: Limited-Tier APIs (Selective)

Tavily (instructor key for demos)
OpenWeatherMap (high daily limit)
Cost: Minimal, managed centrally
Goal: Show production capabilities

Notes on Custom Tool Definitions

For tools marked ❌ (no pre-built LangChain integration), you'll need to create custom tools using the @tool decorator:


xxxxxxxxxx
from langchain.tools import tool

@tool
def my_custom_tool(param1: str, param2: int) -> str:
    """
    Description of what the tool does.
    This docstring becomes the tool description for the LLM.
    """
    # Your implementation here
    result = some_function(param1, param2)
    return result

Then use it with bind_tools():


xxxxxxxxxx
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-4")
tools = [my_custom_tool]
llm_with_tools = llm.bind_tools(tools)

Additional Resources

LangChain Tools Documentation: https://python.langchain.com/docs/integrations/tools
LangGraph Documentation: https://langchain-ai.github.io/langgraph/
Tool Calling Guide: https://python.langchain.com/docs/how_to/tool_calling/
Custom Tools Guide: https://python.langchain.com/docs/how_to/custom_tools/

Last Updated: February 2026