Skip to main content

Tools Overview

Every UltraClaw agent comes equipped with a complete set of built-in tools. You don’t need to install anything — just ask your agent to use them.
Tools are invoked automatically based on what you ask. You can also ask your agent explicitly: “search the web for…”, “run this code”, “read this PDF”, etc.

Built-In Tools

Web Search

Search the internet in real time. Your agent can look up current news, documentation, prices, facts, and anything else that requires up-to-date information.Example: “What’s the latest on OpenAI’s new model?”

Web Browsing

Fetch and read the full content of any web page — articles, docs, PDFs hosted online, product pages. Goes deeper than search results.Example: “Read this article and summarize it: [url]”

Browser Automation

Control a real browser to interact with web pages: fill forms, click buttons, take screenshots, and extract data from pages that require JavaScript.Example: “Go to this page and fill out the contact form for me.”

File Management

Create, read, edit, and organize files in your workspace. Your agent can write documents, save notes, manage project files, and more.Example: “Save a draft of this proposal to my workspace.”

Code Execution

Write and run code to process data, perform calculations, parse files, call APIs, or automate tasks. Supports multiple languages.Example: “Write a Python script to parse this CSV and give me the top 10 rows by revenue.”

Image Analysis

Analyze images you send — read text from screenshots, describe photos, extract data from charts, identify objects, or review documents.Example: “Here’s a screenshot of my error — what’s wrong?”

PDF Analysis

Read and extract content from PDF files. Your agent can summarize, search, and answer questions about PDF documents you share.Example: “Summarize the key points from this contract PDF.”

Sub-Agents

For complex multi-step tasks, your agent can spawn sub-agents to work on parallel workstreams and combine their results.Example: “Research 5 competitors and write a comparison report.”

Already-Documented Capabilities

Some agent capabilities have their own dedicated pages:

How Tool Usage Affects Tokens

Each tool call consumes tokens — the more your agent does, the more tokens it uses. Tool-heavy tasks (like browsing multiple pages, running code, or spawning sub-agents) use more tokens than simple conversations.
Use the /context command to check how many tokens you’ve used in the current session. See Tokens & Usage for more on how credits work.