Join the event trusted by enterprise leaders for nearly two decades. VB Transform brings together the people building real enterprise AI strategy. Learn more
Qodo, the fast growing AI coding startup from Israel focused on code quality, has announced a new collaboration with Google Cloud aimed at advancing the integrity of AI-generated software.
As organizations increasingly rely on artificial intelligence to produce large portions of their codebases, the need for effective oversight and quality assurance tools is growing.
According to Qodo’s CEO Itamar Friedman, AI-generated code is no longer just supplemental — it is becoming foundational to modern development.
“Imagine a future where AI writes all the code; humans won’t be able to review it,” said Friedman. “There must be a system to ensure the values and intentions behind that code remain intact.”
Qodo’s journey has all led up to this point
Before its collaboration with Google Cloud, Qodo had already made notable strides in the enterprise software development space.
Formerly known as Codium, the company rebranded to Qodo in fall 2024 to reflect its growing focus on holistic code quality and AI-driven engineering support.
One of its major breakthroughs came in February 2025 with the release of Qodo-Embed-1-1.5B, a lightweight yet high-performing open-source code embedding model.
The 1.5-billion-parameter model achieved top scores on industry-standard benchmarks like the Code Information Retrieval Benchmark (CoIR), outperforming larger and more resource-intensive offerings from OpenAI and Salesforce. It was designed specifically to help development teams search, retrieve, and contextualize code accurately across vast, multi-language codebases—making it particularly relevant for large-scale enterprise environments.
Qodo has also explored the application of AI agents for regression testing with the release of Qodo Cover in December 2024, a fully autonomous tool that validates whether software changes continue to meet quality and coverage criteria.
Introduced at AWS re:Invent, Qodo Cover demonstrated its capabilities by successfully contributing validated code to Hugging Face’s open-source PyTorch Image Models repository.
The company’s approach emphasizes modular, task-specific AI agents that assist with discrete engineering tasks like test generation, code review, and embedding, rather than pursuing an all-in-one solution. As CEO Itamar Friedman put it, the goal is not just to write more code faster but to ensure that AI-generated code integrates reliably and adheres to best practices in complex systems—a concern especially pressing as AI tools generate an increasing share of enterprise software.
“We see ourselves as responsible for building the machine that observes the machine—verifying not just software practices, but also values,” Friedman said.
Free code review for open source projects
At the center of the partnership is Qodo Merge, an AI-powered code review tool that uses Google’s Gemini models.
Qodo Merge analyzes code for potential bugs, structural issues, and deviations from best practices.
As part of the collaboration, this tool will be offered at no cost to open-source software maintainers, helping ensure quality at the core of widely-used community projects.
Friedman explained the significance of this initiative: “By helping open source projects to maintain the highest quality standards, we’re strengthening the building blocks of the entire development ecosystem while creating intelligent guardrails that let developers move confidently at AI speed.”
Deep integration into a large and growing ecosystem
Qodo’s suite of tools is now natively integrated with Google Cloud, offering users direct access through services like Vertex AI and the Google Cloud Marketplace.
This includes QodoEmbed, Qodo’s compact and high-performing code embedding model, now listed in Vertex AI’s Model Garden.
“Our code embedding model is now in Google’s Model Garden, making deployment much easier for our clients,” Friedman said.
The embedding model enables fast, accurate code search and context retrieval—key features for modern AI coding workflows. It simplifies implementation of retrieval-augmented generation (RAG) pipelines and is optimized for enterprise deployment, including GPU-based infrastructure provided by Google Cloud.
Enhanced code assist built atop Google’s powerful Gemini 2.5 Pro LLM
Another aspect of the integration is the enhancement of Qodo Gen, Qodo’s AI code assistant. By incorporating Google’s Gemini 2.5 Pro model, Qodo Gen now delivers more context-aware and accurate code suggestions, offering developers a smoother AI-assisted development experience.
These improvements also make Qodo’s tools compatible with emerging developer workflows, such as fully automated commit-and-review cycles powered by AI agents working in tandem.
Qodo has designed its platform with flexibility in mind, offering both headless and head-full solutions. This modular approach allows enterprise users to pick and choose components—such as code review, embeddings, or context agents—and integrate them into existing workflows via API, CLI, or native connectors.
“We’re both a headless and headful company, offering modular technology so enterprises can build workflows tailored to their needs,” Friedman said.
This approach appeals particularly to large organizations with complex pipelines. Qodo’s tools are already in use by Fortune 500 companies, including Intuit and NVIDIA. In fact, NVIDIA has showcased Qodo’s capabilities at industry events and detailed how they integrate Qodo’s tools into their own AI workflows.
Supporting startups via Google Cloud Perks
To extend its reach beyond large enterprises, Qodo is participating in Google Cloud’s Startup Perks program.
As part of this initiative, eligible early-stage companies can receive a 50% discount on Qodo’s tools that they are using for non open source, commercial/proprietary products.
The goal is to make enterprise-grade code quality solutions accessible to fast-growing startups looking to scale responsibly.
“We’re enabling startups and enterprises alike to use Qodo seamlessly with Google Cloud, and we’re offering discounts to make that integration even easier,” said Friedman.
This program aligns with Qodo’s broader mission of eliminating the tradeoff between speed and quality in software development—a theme Friedman repeated throughout his remarks about the collaboration.
A commitment to quality above all
As AI tools begin to generate code across all layers of the tech stack—from IDEs and design systems to cloud infrastructure—Qodo sees an urgent need for quality validation tools that operate independently of the code’s origin.
Friedman likened Qodo’s role to that of an observability provider: just as companies like Datadog monitor applications regardless of where they are deployed, Qodo aims to verify code quality regardless of how or where that code is written.
By embedding itself in platforms like Google Cloud, Qodo is positioning its solution as a neutral, cross-platform standard for validating code integrity in a landscape where AI-driven development is rapidly becoming the norm.
Source link