Greptile, a startup that’s building artificial intelligence-based code reviewers to validate human- and AI-generated software, has raised $25 million in an early-stage round of funding as it looks to keep pace with its rivals.

Today’s Series A round was led by Benchmark Capital and saw participation from existing investors Y Combinator, Initialized Capital and the angel investor Cory Levy.

Greptile, which is officially named Tabnam Inc., is looking to solve an increasingly tight bottleneck in the software development lifecycle that has emerged due to the growing popularity of AI coding tools such as Cursor, Devin and Anthropic PBC’s Claude Code. Because so many developers are using these coding agents to accelerate their productivity, the volume of code being produced has mushroomed, and that’s causing problems further down the road.

The issue is that traditional systems for validating code prior to its deployment simply cannot keep up with the energetic pace at which AI coding agents operate. What’s needed, according to Greptile, is yet more AI to enable this AI-powered acceleration. It has created an independent code validation layer that works across code repositories and developer platforms to automate the process of checking code for errors and bugs.

Greptile’s AI code review agent acts like a veteran software developer that never sleeps, constantly analyzing whatever GitHub repositories it’s asked to monitor, checking every new update to ensure that the code works as it should and enforcing company standards. If it finds any bugs or vulnerabilities, it will alert human developers to the problem and also suggest fixes, which can be implemented with a single click.

In a blog post announcing today’s round, Greptile co-founder and Chief Executive Daksh Gupta said adoption of the company’s code review agent has accelerated tremendously. In the last month, it has reviewed more than 500 lines of code for companies such as Substack, Brex Inc., PostHog Inc., Bilt Payments LLC and Y Combinator’s internal software team. During that time, it surfaced more than 180,000 bugs, which were prevented from reaching production.

Greptile v3

The funding announcement was timed to coincide with a major update to Greptile’s code review agent. Today it launched Greptile v3, which represents one of the most significant overhauls of its agentic architecture thus far, Gupta said.

“[It is] as complete rewrite of our core agent architecture, capable of catching three-times more critical bugs than Greptile v2,” he said. “Greptile v3 outperforms every other code reviewer we tested at catching bugs. Some of our customers including Brex have had early access to v3 for a few weeks and were surprised by how much better it is.”

The new version introduces a number of novel features that Gupta said are designed to enhance the quality of the code it reviews and increase the productivity of software teams. For instance, it has gained the ability to analyze the comments of engineers and developers in platforms such as GitHub and GitLab – the reason being, this information is generally kept private by companies, as it provides rich insights into their internal coding standards and best practices. Such comments may provide useful information about how a company handles database transactions and so on, Gupta said.

Also new is the Greptile MCP Server, based on the open-source Model Context Protocol that’s fast becoming the standard way for AI agents to communicate with one another. The Greptile MCP Server allows coding agents such as Devin and Cursor, as well as integrated development platforms to access Greptile’s feedback and organizational rules, so they can improve the code they generate before it’s reviewed.

Additionally, Greptile now integrates with tools such as Notion and Jira to get better context for its code reviews, and gets the ability to detect documentation files such as .corsorrules and CLAUDE.md, which enable it to personalize its code analysis. And for each code fix suggesting it makes, it will create a copyable prompt that developers can use with third-party coding agents to quickly implement them.

No doubt Greptile will be mighty pleased to fill up its war chest with new funds, because even with all of its newly announced capabilities, it faces some tough competition in the nascent AI code review industry. Earlier this month, one of its chief rivals, CodeRabbit Inc., closed on a $60 million Series B round of funding and announced a slew of its own updates. It said it has become the most downloaded application in the GitHub marketplace, allowing it to grow its revenue 10-fold in the last year.

Another key competitor is Graphite, which secured $52 million in funding in March. Graphite benefits from a close partnership with Anthropic and has acquired some big name clients such as Shopify Inc., Snowflake Inc. and Perplexity AI Inc.

Greptile therefore has a lot of work to do if wants to become the top dog of AI code review agents, and will get busy using the money from today’s round to accelerate the development of its platform and make it even better.

Images: Greptile