Home / Founders / Directory / Ian Webster
Ian Webster
Co-founder and CEO · Promptfoo
AI red teaming Open source
Webster turned the testing he needed to ship Discord’s Clyde chatbot into Promptfoo, an open-source tool that lets developers attack their own AI apps and agents before release.
Ian Webster is co-founder and CEO of Promptfoo, an open-source tool for testing and red-teaming apps built on large language models. He is an engineer with a long record of public side projects, and he started Promptfoo while leading AI product teams at Discord.
Asteroids, text messages, and side projects
His personal site lists dozens of projects built over the years, including Asterank, a database for exploring every known asteroid, Textbelt, a free API for sending text messages, QuickChart, which turns a URL into a chart image, and a Kickstarter-funded potato-powered launch to near-space. “I’m an engineer. I’ve always loved building and tinkering,” he said in an a16z conversation.
Shipping Clyde at Discord
Webster joined Discord in 2020 and started its developer platform organization, then moved over to lead early AI products, including Clyde, Discord’s AI chatbot. “We were building this agent and rolling it out to 200 million users,” he said. The team ran into jailbreaks and data-leak risks from a bot that could read private channel history, render images, and search the web, and the checks they wrote went into the first version of Promptfoo.
Building Promptfoo
Promptfoo began as an open-source command-line tool for comparing prompts and running evaluations on a developer’s own machine. Webster and Michael D’Angelo turned it into a company in 2024, focused on automated red teaming that probes apps for prompt injection, jailbreaks, and dangerous combinations of agent tools. The company says its tools are used at more than a quarter of Fortune 500 companies.
In March 2026 OpenAI announced an agreement to acquire Promptfoo and fold its security testing into OpenAI Frontier, its platform for enterprise agents, while keeping the open-source project going. “As AI agents become more connected to real data and systems, securing and validating them is more challenging and important than ever,” Webster said in the announcement.
