Report Abuse

Basic Information

Notte is an open source full‚Äëstack framework for building, testing, deploying, and scaling AI agents that interact with the web. It combines traditional deterministic scripting with AI-driven reasoning so developers can script routine interactions and use models only where adaptive decision making is needed. The project supplies both a local Python SDK and a hosted API service that manages browser sessions, and it targets reliability, speed, cost efficiency, and scale for web automation tasks. Core capabilities exposed in the README include running agent tasks against websites, producing structured outputs via Pydantic models, Playwright‚Äëcompatible site interactions, stealth browser sessions with anti‚Äëdetection features, and endpoints for scraping and session orchestration. The repo is suitable for teams that need programmatic web automation with integrated LLM reasoning and enterprise features like credential vaults and digital personas.

Links

Categorization

App Details

Features
Notte documents a set of developer-facing features for web automation agents. It supports running agents locally or via a hosted API and a Python SDK. Agents can return structured output by accepting Pydantic models as response formats. The framework exposes Playwright‚Äëstyle primitives for observing and executing page actions and a scraping endpoint for automated data extraction. Session features include stealth capabilities with CAPTCHA solving and proxy support, cookie and auth session handling, CDP browser compatibility, and file upload/download storage scoped to sessions. Higher level tools include an Agent Vault for secure credentials, Agent Personas that provide emails, phone numbers, and automated 2FA handling, and hybrid workflows that mix scripts and LLM reasoning to reduce costs and improve reliability.
Use Cases
Notte helps teams automate complex web tasks by combining deterministic scripting with LLM reasoning to reduce API usage and increase task reliability. Structured output support ensures extracted data conforms to developer schemas, simplifying downstream processing. Session features like built‚Äëin CAPTCHA solving, proxy configuration, cookie management, and CDP compatibility improve success rates against real websites. Vaults and persona tools streamline authenticated flows and account creation while keeping secrets managed. The hosted SDK and session API let teams prototype locally and then switch to managed browser sessions for scale. The scraping endpoint and hybrid workflow examples make it easier to extract, format, and act on web data in a cost‚Äëefficient, reproducible way.

Please fill the required fields*