Periscope Founders Launch Macroscope AI for Code Insights

Periscope founders return with a developer-focused AI

Kayvon Beykpour and Joe Bernstein, the team behind Periscope, are back with Macroscope — an AI-powered understanding engine built to reduce the friction of large codebases. Joined by ML veteran Rob Bishop, the founders aim to give engineers and product leaders a clear, concise view of what changed, what broke, and what teams actually shipped.

Macroscope integrates with GitHub Cloud and optional tools like Slack, Linear, and JIRA. It performs “code walking” using Abstract Syntax Trees (ASTs) to build structural context and then layers large language models (LLMs) on top to generate human-friendly summaries and findings. The result: fewer status-check meetings and faster onboarding for stakeholders who need answers without interrupting engineers.

What teams can get from Macroscope:

• Automatic PR summaries and context for reviewers
• Bug discovery in pull requests using AST-aware analysis
• Real-time product and productivity summaries for leaders

In benchmarks shared by the company, Macroscope caught about 5% more real-world bugs than the next-best tool and generated 75% fewer comments, reducing review noise. Pricing starts at $30 per active developer per month with enterprise options and requires GitHub Cloud.

Macroscope sits in a crowded code-review and developer tooling market — competitors include CodeRabbit, Cursor Bugbot, and Graphite — but its AST-driven approach combined with LLMs aims to capture both syntactic precision and natural-language clarity. Early customers include startups and scale-ups such as XMTP, United Masters, and A24 Labs.

Why this matters: As engineering organizations scale, visibility into who built what and why becomes expensive in time and coordination. Macroscope promises an automated middle layer that translates code changes into operational intelligence — think of it as a radar that points leaders and devs to meaningful change instead of noise.

There are still trade-offs to evaluate: security and privacy of code access, reliance on GitHub Cloud, and how well LLM-generated summaries align with a team’s internal context and standards. Large organizations will want pilot data on false positives, missed bugs, and measurable time savings before committing broadly.

Financially, Macroscope closed a $30 million Series A led by Lightspeed in July and has raised $40 million to date. The team of about 20 is aiming to expand enterprise integrations and refine its AST+LLM models to scale across languages and architectures.

For engineering and product leaders, Macroscope is another example of AI shifting from assistant-like helpers to tooling that embeds directly into developer workflows. The question for buyers: will this reduce context switching enough to justify the cost and trust of deep repo access? Early benchmarks look promising, but careful pilots will be key.