AI & TechArtificial IntelligenceNewswireStartupsTechnology

Unlock Your Codebase with Macroscope: AI-Powered Bug Fixing

▼ Summary

– The founders of Periscope have launched Macroscope, an AI system for developers and product leaders that summarizes codebase updates and catches bugs.
– Macroscope was co-founded in July 2023 by Kayvon Beykpour, Joe Bernstein, and Rob Bishop, all with prior successful startup exits.
– The product integrates with tools like GitHub, Slack, and JIRA to analyze code and provide insights, aiming to save engineers time and reduce meetings.
– In benchmarks, Macroscope caught 5% more bugs than the next-best tool and generated 75% fewer comments.
– The startup has raised $40 million total, including a $30 million Series A led by Lightspeed, and charges $30 per developer per month.

Macroscope offers a powerful AI-driven solution for developers and product leaders seeking to streamline codebase management and enhance bug detection. Founded by former Twitter head of product Kayvon Beykpour, this new startup introduces an intelligent system designed to save engineering teams significant time and effort. The platform provides automated summaries of code changes, identifies potential bugs, and answers natural language queries about development progress.

Beykpour, now CEO of Macroscope, co-founded the company in July 2023 alongside Joe Bernstein, a longtime collaborator from their earlier ventures including Periscope and Terriblyclever. They are joined by Rob Bishop, who brings expertise from his computer vision and machine learning background. Together, they built a tool they describe as an “AI-powered understanding engine”, something they wish had been available during their previous entrepreneurial experiences.

Many engineering teams currently rely on a patchwork of project management tools such as JIRA, Linear, and spreadsheets, often spending excessive time in meetings rather than actual coding. Macroscope aims to change that by offering a unified platform that automates oversight and insight.

Beykpour shared that his own frustrations at both startups and large organizations like Twitter inspired the creation of Macroscope. “Understanding what everyone was working on, especially in a company with thousands of engineers, consumed most of my time, and it was my least favorite part of the job,” he recalled.

To get started, users install Macroscope’s GitHub application, granting it access to their codebase. Optional integrations with Slack, Linear, and JIRA are also available. The system employs a technique called code walking, which leverages the Abstract Syntax Tree (AST) to structurally analyze programming code. This analysis, combined with large language models, allows Macroscope to gather deep contextual understanding of how the code operates.

Once integrated, engineers can use Macroscope to detect bugs in pull requests, summarize PR content, track codebase evolution, and conduct code-related research. Product leaders benefit from real-time updates on development activity, productivity analytics, and answers to plain-language questions about the product or engineering priorities. This helps leadership make informed decisions about resource allocation without interrupting workflow.

You can ask questions in natural language, no matter your technical skill level,” Beykpour emphasized. “This is incredibly useful for learning about the codebase without pulling senior engineers away from their work. If you’re a CEO and want to know what got done this week, you can either ask Macroscope or interrupt your team. One option is far more costly.”

While several tools operate in the code review space, such as CodeRabbit, Cursor Bugbot, and Greptile, Macroscope claims a competitive edge. Internal testing on over 100 real-world bugs showed that Macroscope identified 5% more issues than the next leading tool while generating 75% fewer comments. These results have been shared publicly in a detailed benchmark report.

Pricing starts at $30 per active developer monthly, with a minimum of five seats. Enterprise plans and custom integrations are available for larger organizations. The software requires GitHub Cloud. Early adopters include companies like XMTP, United Masters, Bilt, Class.com, and A24 Labs.

Based in San Francisco, Macroscope employs a team of 20 and recently secured $30 million in a Series A round led by Lightspeed’s Michael Mignano. Additional investors include Adverb, Thrive Capital, and Google Ventures. Total funding to date amounts to $40 million.

(Source: TechCrunch)

Topics

ai startup 95% code summarization 90% bug detection 88% developer tools 85% github integration 80% productivity insights 78% natural language queries 75% code review 72% abstract syntax tree 70% large language models 68%