Meet Macroscope: an AI-powered solution designed to help you comprehend your codebase and resolve bugs

The creators who once built and sold the live video platform Periscope to Twitter have returned with a fresh venture—this time, they're diving into artificial intelligence.

On Wednesday, Kayvon Beykpour, the former product chief at Twitter, revealed Macroscope, a new AI platform tailored for developers and product managers. The tool summarizes codebase updates, detects bugs, and handles various other coding tasks.

Beykpour, now serving as CEO, founded the company in July 2023 along with his long-time friend Joe Bernstein, who also worked at Periscope and previously co-founded Terriblyclever, which was acquired by Blackboard in 2009. They brought in Rob Bishop as a co-founder, who previously sold his computer vision and AI company Magic Pony Technology to Twitter in 2016.

Macroscope is positioned as an “AI-driven understanding engine” that helps engineers reclaim their time—a tool the founders say they wished they’d had during their earlier startups.

According to Beykpour, engineers today rely on a mix of platforms like JIRA, Linear, and spreadsheets to manage tasks, often spending excessive time in meetings rather than building. Macroscope aims to solve this problem.

“I really experienced this pain firsthand…at every company, whether it was our own startups or large public companies like Twitter, we faced this issue the hard way,” Beykpour explained to TechCrunch in an interview.

“Understanding what everyone was working on, especially in a company as big as Twitter with thousands of engineers, became the core of my job—and honestly, it was the part I liked the least while leading product at Twitter,” he added.

To tackle these challenges, Macroscope users start by installing its GitHub app, which allows access to their codebase. Additional integrations, such as Slack, Linear, and JIRA, can be added if desired. The system then automatically analyzes code changes and tracks updates.

The platform utilizes a process called code walking, which leverages the Abstract Syntax Tree (AST)—a framework representing code structure—to collect important context about the codebase. This is combined with large language models (LLMs) to enhance understanding.

After setup, engineers can use Macroscope to spot bugs in pull requests, receive concise PR summaries, track codebase changes, and submit code-related queries. Product managers, meanwhile, can get instant product updates, productivity analytics, natural language answers about development activities, and more—helping them understand how engineering resources are allocated.

“You can ask questions in plain English, regardless of your technical background,” Beykpour pointed out. “This is great if you want to learn about the codebase without interrupting a senior engineer. Incredibly helpful. For executives, if you want to know, ‘What did we accomplish this week?’ you can either ask Macroscope or interrupt your team. One is much more efficient than the other.”

Although Macroscope doesn’t have a single direct competitor with the same feature set, it does operate in the code review field—where tools like CodeRabbit, Cursor Bugbot, Graphite Diamond, and Greptile also work. In internal tests on over 100 real-world bugs, Macroscope identified 5% more bugs than its nearest competitor and generated 75% fewer comments. (The company published these results in a blog post.)

Macroscope is priced at $30 per active developer per month, with a minimum of five users. Enterprise packages and custom integrations are available for larger organizations. The product requires GitHub Cloud. Before its public release, startups and established companies such as XMTP, Things, United Masters, Bilt, Class.com, Seed.com, ParkHub, and A24 Labs have already adopted the tool.

Based in San Francisco, Macroscope’s team consists of 20 people and has secured $30 million in Series A funding, led by Michael Mignano at Lightspeed in July. Other backers include Adverb, Thrive Capital, and Google Ventures, bringing the total raised to $40 million.