What We Built at Volt's First Internal Hackathon

The Projects

Clippy

Clippy (built by Migue and Mati) is an AI sidekick that lives inside the Volt Chrome extension as a right-side drawer.

Under the hood, it's a tool-calling agent built on OpenAI. It can read your messages, find and open specific chats, draft and send messages across multiple conversations, archive, pin, mute, schedule, create groups, and run intro flows in the right language.

You can also connect your Google account and it also pulls from Calendar, Drive, and Contacts via Composio.

The mascot is the famous paperclip from Microsoft Offce '97.

Before every response, it delivers exactly one snarky one-liner about what you asked.

WhipIt

WhipIt kills the friction of juggling tickets across multiple repos: branch-switching churn, stashed half-finished work, an AI that relearns your project every session.

Santi Vilar built a control center above your real repos the code lives as canonical clones, and every task gets its own git worktree, so several Linear issues stay in flight with zero churn.

A few slash commands run the loop: pull this cycle's issues, sort them ready / needs-clarification / blocked, spin up worktrees on Linear's branch names so PRs auto-link, tear them down once merged.

The operational knowledge lives in instructions the assistant reads every session, so the dev process becomes versioned muscle memory instead of context you re-explain each morning.

WDS Lab

WhatsApp design system (WDS) components are hard to review inside the real WhatsApp runtime. You'd normally build a one-off test page just to check a button state.

WDS Lab, built by Juli Caruso and Santi Santana, is a fullscreen developer-only drawer inside Volt that lets the team preview component states, prop combinations, and native styling in context, with no throwaway test pages needed.

AI Inbox

AI Inbox organizes your chats into 6 labels using AI. It queries your last 100 chats and assigns the best tag to each conversation.

It was built by Manu Achinelli and Manu Mao, they handle 100s of conversations each week and needed a dynamic way to organize their chats. They took inspiration from Fyxer AI.

Pepe Roni

Tom Mercado built an E2E Playwright Testing System, because testing a Chrome extension that sits on top of WhatsApp Web is hard.

The extension has to run inside a real browser with a real WhatsApp session, reacting to real UI events.

The suite Tom built runs via a dedicated playwright.e2e.config.mjs and starts the extension runtime through a global setup script.

Instead of launching a fresh browser each time, it reuses a persistent local Chrome profile that already has the extension installed and WhatsApp connected.

What's covered

Four core feature areas now have automated coverage:

- Private lists — creating, editing, and deleting lists, plus assertion scenarios for edge cases
- Tasks from messages — the flow of turning a WhatsApp message into a task
- Workspace membership & list sync — simulated multi-member workspace flows with list state validation
- Integrations — coverage for the integrations panel and connection flows

What a Hackathon Surfaces

All 5 projects came from problems the team hits daily. Clippy addresses the gap between AI agents and WhatsApp. WhipIt removes the friction of context-switching across three repos. WDS Lab stops the "just deploy it to check" cycle for UI work. Labels is something users keep asking for.