Introduction
Agentlet augments existing web applications with form automation, data extraction, AI integration, and UX improvements. It deploys via lightweight bookmarklets, browser extensions, or native integration, without backend changes.
This removes the immediate need for backend modifications, complex deployments, or lengthy development cycles, letting teams rapidly enhance user workflows.
Agentlet offers three deployment approaches: bookmarklets for instant deployment, browser extensions for enhanced capabilities, and native integration for applications that can embed the framework directly. Each approach delivers the same core augmentation capabilities while adapting to different technical environments and constraints.
Where to go next
Section titled “Where to go next”- Getting started: clone agentlet-core, run the built-in demo, scaffold your own agentlet, or set it up manually.
- Concepts: the philosophy behind agentlet, its architecture, deployment modes, and security model.
- Guides: task-oriented documentation for forms, tables, AI, authentication, environment variables, dialogs, shadow DOM, the mount API, TypeScript, script injection, and z-index layering.
- Reference: the full
window.agentletpublic API, organized by namespace. - Contributing: commit message rules and the documentation style guide for this project.
What agentlet is for
Section titled “What agentlet is for”Agentlets are tactical tools: they let teams rapidly improve processes, test ideas, or temporarily enhance applications while waiting for more robust, long-term solutions. They favor speed over perfection, and are not meant to replace deeper system integrations.
The agentlet ecosystem includes the core framework (agentlet-core), example implementations, and agentlet-designer, a dedicated agentlet for building custom agentlets.
Source: agentlet-core README.md, sections “Agentlet” and “The agentlet approach”.