Problem

Caddy addresses the fragmentation and friction of modern knowledge work. Work is spread across Slack, Linear, Gmail, calendars, and other tools, forcing people to jump between tabs and repeatedly translate intent into low-value clicks and typing. The resulting context switching, lost focus, and hours spent on repetitive work are the core economic pain, although the available evidence does not quantify the cost in dollars.

The main use case is a busy founder, operator, or productivity power user saying something like “follow up on this” and having Caddy understand the relevant screen context and complete the task. Examples include creating a Linear ticket, sending a Slack message, scheduling a meeting, or handling a task that the user otherwise would postpone.

Product / Service

Caddy is a personal AI assistant delivered through messaging and iMessage. The current product page says there is no app to download: users can text, send voice memos, or add Caddy to group chats, while Caddy works in the background across Gmail, Calendar, Slack, Linear, and other tools. It can proactively text first when something matters, monitor calendars, chats, and tasks, nudge the user, or take action on the user’s behalf.

Caddy also personalizes itself through usage, learning how a person works, communicates, and prioritizes. YC’s launch description presents the underlying interaction model as a desktop voice interface with an action mode for operating applications and a dictation mode for replacing typing across the computer. The benefit is a more natural, proactive layer over existing work software that turns spoken or text intent into completed work rather than merely generating generic answers.

Market

Caddy competes in the emerging market for personal AI assistants, agentic work automation, and voice interfaces for computers. Its adjacent alternatives include general voice assistants such as Siri and Alexa, which Caddy’s own description characterizes as useful for trivia but limited at doing real work, as well as cross-application assistants such as Vellum and Morgen that cover combinations of email, calendars, Slack, tasks, and messages. Caddy’s intended differentiation is proactive behavior, screen and workflow context, cross-app execution, and personalized work artifacts.

The company was founded in 2025, is an active Fall 2025 Y Combinator company, and is listed with a five-person team. Its public materials have promoted a launch and early access or waitlist, while the product site now presents a usable messaging experience. No customer count, revenue figure, or other quantified adoption metric appears in the available evidence, so Caddy is best characterized as an early-stage company in launch or early-access phase, with no public evidence here that it has reached meaningful commercial scale.