Browser/Computer-Use Agents
These agents don't just call APIs — they drive a real browser or desktop, clicking, typing, and navigating UIs the way a human would. Useful for testing, scraping, and automating workflows that have no API to begin with.
-
SkyvernOpen-source browser automation agent that uses computer vision and LLMs to fill out forms and navigate workflows on websites it has never seen before, without hand-written selectors.
-
MultiOnBrowser agent platform and API for building autonomous web agents that can shop, book, research, and complete multi-step tasks across live websites on a user's behalf.
-
OctoMindAI QA platform that explores a web app's clickable elements and flows, then generates, runs, and self-heals Playwright tests with no coding required.
-
Playwright MCP
Microsoft's MCP server that lets LLM agents drive a real browser through Playwright using accessibility snapshots instead of screenshots.
-
Stagehand
Open-source browser automation SDK from Browserbase that mixes Playwright code with natural-language act, extract and observe primitives.
-
Steel
Open-source headless browser API for AI agents with session management, proxies and CAPTCHA handling, available self-hosted or as a cloud service.
-
Hyperbrowser
Cloud browser infrastructure for AI agents and scrapers, offering scalable sessions, stealth and CAPTCHA solving with Playwright and Puppeteer support.
-
NovaAI test automation platform that generates end-to-end tests from real user behavior and updates them automatically as the app changes.
-
Checksum AIContinuous testing platform whose AI agents generate and maintain end-to-end tests from real user sessions, self-healing around 70% of failures.
-
KushoAIAutonomous test-suite generator for APIs and UIs. Feed it an OpenAPI spec, Postman collection, or cURL command, and it produces a full test suite.
-
MeticulousAutomated end-to-end screenshot testing that records real sessions, then replays them deterministically against every pull request to catch visual regressions.
-
Browserbase
Cloud infrastructure for running headless browsers at scale for AI agents, handling stealth fingerprinting, residential proxies, and CAPTCHA solving.
-
Browser Use
Open-source framework that wraps Playwright so an LLM can control a browser from natural-language instructions. Around 117,000 GitHub stars.