Cloud browser infrastructure and a Playwright-based framework for running web automations with AI agents, with stealth, CAPTCHA solving and proxies.
tinyfish-io/agentql
A toolkit for extracting data and automating live websites with an AI-powered query language that works with Playwright, plus SDKs and a REST API.
- Stars
- 1,491
- Forks
- 179
- Open issues
- 17
- Last commit
- 6 hours ago
What is it?
AgentQL connects LLMs and AI agents to web pages so they can extract data and automate workflows on live sites. You describe the elements or data you want in a natural language query, and AgentQL locates them on the page, including content behind authentication or generated dynamically. Queries can define the structure of the output and apply transforms during extraction.
The project ships Python and JavaScript SDKs that integrate with Playwright, a REST API for running queries without an SDK, and a browser extension for debugging and refining queries on live pages. It also offers integrations with agent frameworks such as LangChain and Zapier, along with an MCP server.
Because its selectors are based on page content rather than fixed DOM paths, the same query can work across similar sites and keep working as the UI changes.
Key features
- 01Natural language queries locate elements and data on any page, including behind authentication.
- 02Python and JavaScript SDKs integrate with Playwright.
- 03Query shapes define structured output and can apply transforms during extraction.
- 04The same query can be reused across sites with similar content.
- 05A REST API and browser debugger extension are included.
Use cases
- Scraping product prices or listings from e-commerce sites.
- Automating logins, form submissions, and cookie dialog handling in browser scripts.
- Giving an LLM agent the ability to read and act on web pages through MCP or LangChain.
Quick start
- 1
Open the quick start guide and install the Python SDK from PyPI
Topics
FAQ
- What is agentql?
- A toolkit for extracting data and automating live websites with an AI-powered query language that works with Playwright, plus SDKs and a REST API. AgentQL connects LLMs and AI agents to web pages so they can extract data and automate workflows on live sites. You describe the elements or data you want in a natural language query, and AgentQL locates them on the page, including content behind authentication or generated dynamically. Queries can define the structure of the output and apply transforms during extraction.
- Is agentql free to use? What license does it have?
- agentql is open source under the MIT license. Check the license terms before using it in commercial products.
- Is agentql still maintained?
- Actively maintained. The last commit to agentql was on Oct 22, 2026.
- What can I use agentql for?
- Scraping product prices or listings from e-commerce sites. Automating logins, form submissions, and cookie dialog handling in browser scripts. Giving an LLM agent the ability to read and act on web pages through MCP or LangChain.
- How do I install agentql?
- Open the quick start guide and install the Python SDK from PyPI
- Where can I get agentql?
- The source code and installation instructions are on GitHub: https://github.com/tinyfish-io/agentql
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source browser automation that uses LLMs and computer vision to run web workflows on any site, extending Playwright with AI commands.
Open-source browser API for AI agents and apps, running a managed Chrome sandbox you can drive with Puppeteer, Playwright, or Selenium.
TypeScript, Python and Go SDK for browser agents that drives Chrome with natural-language act, observe and extract calls alongside Playwright-style APIs.
Open-source Python library for building AI agents that control a real browser, with optional Browser Use Cloud stealth browsers and hosted agent API.
An open-source autonomous research agent that plans questions, gathers web and local sources, and writes cited reports with any LLM provider.