dotdotduck
Web agent that operates the page, not a sidebar chatbot
dotdotduck is an embeddable SDK that enhances web applications by providing intuitive interaction capabilities. Key features include:
• Unified Command Palette for feature discovery
• Web Agent for multi-step task completion directly on the page
• Inline Text Operations for context-aware content manipulation
• Voice Input for natural language interaction
• Structured Intent Capture for user behavior analysis
This SDK allows users to find features, execute tasks, and generate insights within their workflow without navigating away from the application. The Command Palette, accessible via Ctrl+K, consolidates all registered commands and an "Ask" function into a single entry point, streamlining user interaction. The Web Agent operates directly within the user's browser tab, performing sequences of actions like navigation, clicks, and form fills, narrated in real-time. This provides a guided experience for complex operations, with user feedback integrated when decisions are required.
Further enhancing user productivity, the Inline Text Operations allow users to select text within any input field and perform actions such as translation, summarization, or rewriting directly in place. This eliminates the need to copy-paste or switch to external tools. Every interaction within dotdotduck generates a structured intent signal, providing application teams with valuable data to understand user needs, feature adoption, and overall product engagement, fostering continuous improvement.
Designed for product teams and developers, dotdotduck integrates quickly into existing web applications, offering extensive customization options through CSS variables and a Skill SDK for custom scripts, prompts, and actions. It supports various voice providers and is ideal for enhancing user experience in SaaS products, internal tools, and any web-based platform where user engagement and feature discoverability are critical.