AI assistant for web automation

Speck is an AI assistant that helps knowledge workers automate their repetitive web workflows. Our customers use us to save hours on data entry, CRM updates, and LinkedIn outreach. Since our launch 3 weeks ago, we've published Speck on the Chrome Web Store and acquired 17 paying users. Lucas helped build Sweep, an AI junior developer. Raghav won 7 hackathons with LLM agents. ‎ Join our community: https://discord.gg/speck

Team Size:2
Location:San Francisco
Group Partner:Diana Hu

Active Founders

Lucas Jaggernauth

Ex-Roblox engineer. Founding Engineer of Sweep (YC S23), an AI junior developer with 7k GitHub stars. At Roblox, I built multilingual autocomplete for 160 million users. I was a computer vision researcher for the BRIAR government project. Graduated Summa Cumme Laude from UT Dallas with a double major in CS and cognitive science.

Lucas Jaggernauth
Lucas Jaggernauth

Raghav Pillai

Ex-Anduril and Ex-Amazon engineer. At Anduril, I led efforts for telemetry, realtime video and health for WISP. At Amazon, I helped build Amazon Astro, a household robot. I previously did AV research and built the software orchestration for a self-driving car. I've won 7 hackathons with LLM agents, like HackRice and HackHarvard.

Raghav Pillai
Raghav Pillai

Company Launches

tl;dr: Speck is a browser extension that makes automating repetitive web tasks easy. Our customers use Speck to extract data, perform LinkedIn outreach, collect the newest research papers, gather data about social media groups, and more!

Hey everyone! We’re Lucas and Raghav, the founding team behind Speck.

🤔 The Problem

People spend precious time doing menial and repetitive browser-based tasks, such as entering data into CRMs, posting multiple job listings for recruiting, reaching out to contacts on LinkedIn, extracting data from websites, and other similar chores.

These tasks are irritating for most knowledge workers, yet they do not have any way to simplify their workflow. Low-code solutions like UiPath cost tens of thousands of dollars and require a dedicated automations team. On the other hand, executing web-based workflows is outside of the scope of workflow tools like Zapier.

💡 The Solution

With Speck, you can automate any workflow easily without having any coding knowledge. Our browser extension allows you to record workflows just by doing them. Speck automatically translates your workflows into repeatable steps, making it simple for anyone to create web-based automations and data extraction processes

By keeping track of intent, the browser automations will repair themselves automatically when something goes wrong! These workflows can be executed live or scheduled asynchronously on a remote browser. Now, our users can spend less time doing menial chores and more time focusing on higher-impact tasks.

Since pivoting 2 weeks ago and launching our beta, we’ve gotten our extension published on the Chrome Web Store and onboarded our first paying customers.

Making an automation takes less than 60 seconds.

Step 1 - Record your workflow

Step 2 - Extract Text

Step 3 - Run

And this is just the start! Together with our users, we’re making progress towards our vision of an autonomous LLM agent capable of generating complex automations - one workflow at a time.

🏆 The Team

We’re both CS from UT Dallas. We met at a hackathon 2 years ago and have been friends since. We both come from top tech companies and came up with the idea for Speck after we grew tired of doing the same actions over and over again on websites. From left to right:

Raghav (CTO) - Ex-Anduril and Ex-Amazon engineer. At Anduril, he created LLM tooling and device monitoring, used by 140 employees. He was an early engineer on Amazon’s Consumer Robotics division, working on Amazon Astro. He worked in AV research building a modular FSD stack. He also loves going to hackathons and won 7, including HackHarvard and HackRice, with LLM agents.

Lucas (CEO) - Ex-Roblox engineer where he built multilingual autocomplete for 160 million users. He was a founding engineer of Sweep, an autonomous LLM coding agent with 7k stars. Computer vision researcher for the BRIAR government project. He graduated Summa Cum Laude with a double major in computer science and cognitive science.