Run Browser Automation Agents with Notte
Build and run browser agents that combine deterministic Playwright-style actions with natural-language AI steps, hosted sessions, vault-backed credentials, and structured outputs.
npx skills add agentskillexchange/skills --skill run-browser-automation-agents-with-notte
Use Notte to create browser automation agents that open browser sessions, inspect page state, scrape data, execute actions, and combine scripted browser primitives with LLM-driven steps for websites that need flexible automation.
What this skill actually does
What the agent/operator does
When to invoke
Invoke this skill when a user needs an agent to complete multi-step web tasks, extract structured data from dynamic sites, test or operate browser workflows, or use managed browser infrastructure with optional CAPTCHA, proxy, credential vault, and digital persona support.
Scope boundary
This is not a generic browser automation listing. Keep the work scoped to a concrete Notte session and task: choose local or hosted execution, define the browser task and output schema, attach safe credentials only when needed, run the agent, and return artifacts or structured results.
Inputs and prerequisites: Python environment, Notte package, browser dependencies such as Chromium via patchright, LLM API key for local use, optional NOTTE_API_KEY for hosted browser sessions.
Setup notes: Install with pip install notte , then run patchright install –with-deps chromium for local browser execution. For hosted sessions, create a Notte API key and set NOTTE_API_KEY before using the Notte client.
Source and verification boundary: use https://www.notte.cc/ as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.