A Look Back at 2024: Qodo’s Key Achievements and Updates

As we close out 2024, we want to share some highlights from Qodo’s journey over the past year and how we’ve come closer to achieving our broader vision of quality-first AI coding. Here’s a recap: 

A new name, $40M funding

In 2024, Qodo grew to become a comprehensive multi-agent AI platform built to enhance code quality across all aspects of software development, from writing code, to testing and reviewing it. To better reflect our growing offering and vision, we embraced a new identity as Qodo. It’s a name that embodies our commitment to placing quality at the core of AI-assisted software development. 

Along with our new name, Qodo achieved a major milestone by securing a $40 million Series A funding round. This investment is supporting the expansion of our AI agents and the enhancement of our platform capabilities to address the complexities of modern software development. Our focus remains steadfast on code quality and delivering advanced tools that empower developers to build resilient, maintainable applications and services. Key innovations have focused on improving test generation, advancing deeper context-awareness, and empowering AI to reason through complex coding challenges, all with the goal of enhancing efficiency and quality throughout the software development lifecycle.

Celebrating adoption milestones

In 2024, Qodo grew significantly, achieving important adoption milestones. Qodo Gen, our IDE agent for code and test generation, reached over 1 million installations (+397K JetBrains / +512K VS Code). Qodo Merge, our code review agent, now processes more than 20,000 pull requests daily. We’re also excited to have added several Fortune 500 companies to our growing list of customers.

Controllable, context-aware AI coding for the enterprise

One of the year’s major announcements for us at Qodo was the launch of our enterprise code integrity platform, helping to address the unique challenges of AI coding at scale. The platform focuses on embedding quality-first generative AI capabilities into coding workflows while tailoring solutions to enterprise-specific needs. 

At the core of the platform is advanced retrieval-augmented generation (RAG), enabling AI coding tailored to specific organizational best practices and coding standards. Organizations can connect their codebases—whether hosted in the cloud or on-premise—for indexing, tagging, and retrieving relevant contextual information needed to generate high-quality code, tests, or review code. Complementing this deep understanding of code context, Qodo’s enterprise platform also features a sophisticated database of best practices. This dynamic system automatically learns and reinforces an enterprise’s specific coding standards, adapting to the unique requirements and preferences of each organization. As development teams use the platform, the system continually refines its understanding of the company’s coding conventions, architectural patterns, and quality benchmarks.

Key features of the enterprise platform include:

• Dynamic learning of organizational coding practices: The platform adapts over time to the company’s evolving conventions and quality benchmarks.
• Optimize context-awareness: Control what’s indexed and tag important files or repos to help AI deliver more relevant and higher-quality code suggestions
• Flexible deployment options: Cloud, on-prem, and air-gapped deployments to meet enterprise-grade security requirements.
• Customizable language model integration: Support for Qodo’s proprietary AI models or integration with the enterprise’s preferred LLMs.

Streamlining code reviews with the Qodo Merge Chrome Extension

One of Qodo’s most popular agents, Qodo Merge, was released as a free Chrome extension (only for public repos) to streamline and enhance code review workflows. Integrated with GitHub, the extension enables AI-chat directly in the pull request and helps to automate code review workflows, enabling developers to focus on critical issues and reduce manual review effort.

Key Features:

• AI-powered chat in PRs: Ask targeted questions about specific code changes and receive context-aware answers within the pull request comment section.
• Automated review workflows: Instantly review pull requests and identify potential bugs, security vulnerabilities, and adherence to best practices with the /review command.
• Code improvement suggestions: Use the /improve command to get actionable recommendations for optimizing code efficiency, robustness, and maintainability.
• Auto-generate PR descriptions: Save time with the /describe command to create comprehensive PR descriptions, including summaries, labels, and step-by-step walkthroughs.
• Private chat sessions: Your conversations remain private and are not visible to other users.
• Streamlined configuration export: Effortlessly export your preferred PR-Agent settings to a configuration file, which can be applied at either the repository or organizational level for seamless team-wide adoption.

The Qodo Merge Chrome extension is free to use on open source repositories, with a paid version available for private repos. 

Qodo Cover: automating test coverage  

During AWS re:Invent, we announced Qodo Cover to address a key challenge in software development: improving meaningful test coverage. An intelligent agent designed to autonomously generate and extend test suites, Qodo Cover integrates into development workflows via CLI or as a GitHub Action, making it easy for teams to improve code coverage without manual intervention. The tool analyzes source code, identifies gaps in coverage, and generates meaningful, high-quality tests tailored to the application’s needs.

By automating the generation and enhancement of test suites, Qodo Cover helps development teams maintain reliable and efficient testing practices, obtain meaningful test coverage, and reduce the risk of undetected issues.

Key Features

• Integration with CI Pipelines: Qodo Cover can be triggered automatically as part of CI workflows, enhancing test coverage during pull requests or scheduled runs.
• Context-Aware Test Generation: The tool analyzes the entire codebase to create tests that align with application logic and project requirements.
• Quality-Focused Testing: Tests generated by Qodo Cover are validated to ensure they add meaningful coverage, avoiding redundancy or unnecessary overhead.
• Detailed Coverage Reporting: Developers can track progress and assess the value of new tests using comprehensive reports.

Qodo Cover made practical contributions to well-known projects in 2024. For instance, it autonomously generated 15 new tests for Hugging Face’s PyTorch Image Models repository, adding meaningful coverage for 168 lines of code. Its ability to integrate into complex projects demonstrated its effectiveness in improving code quality without significant manual effort.

Read more about Qodo Cover in this blog post.

AlphaCodium: prompt engineering to flow engineering 

In 2024, we published our seminal AlphaCodium research—a flow engineering framework that enhances the performance of LLMs in solving complex coding challenges and brings AI closer to system 2 thinking.  AlphaCodium employs a test-based, multi-stage, iterative process that emphasizes continuous improvement through iteration, moving beyond traditional prompt engineering to flow engineering. 

Key capabilities of AlphaCodium

• Iterative Code Refinement: AlphaCodium utilizes a structured process involving code generation, execution, testing, and refinement. This approach ensures that the system arrives at a fully validated solution, improving accuracy and reliability.
  Enhanced Problem Understanding: By generating additional data such as problem reflections and test reasoning, AlphaCodium aids LLMs in better comprehending complex problems, leading to more effective solutions.
• AI-Generated Test Enrichment: The system enriches existing public test cases with AI-generated tests, introducing additional layers of scrutiny to validate the robustness of the solutions.

AlphaCodium demonstrated significant performance enhancements when applied to challenging code generation datasets like CodeContests, which includes competitive programming problems from platforms such as Codeforces. When combined with OpenAI’s o1 model, AlphaCodium elevated performance significantly, achieving better results compared to o1 alone and even outperforming o1-ioi, a model further trained specifically for programming skills.

The principles and best practices developed through AlphaCodium are broadly applicable to general code generation tasks. By focusing on a flow that emphasizes problem understanding, iterative refinement, and rigorous testing, AlphaCodium provides a framework that can be applied to various LLMs pre-trained to support coding tasks.

For more detailed insights and access to the open-source implementation, you can refer to the official AlphaCodium GitHub repository and the research paper.

Qodo in the News

In 2024, Qodo achieved several noteworthy milestones, reflecting its progress and growing recognition in the industry. Here are the highlights:

• Inclusion in the Gartner Magic Quadrant
  Despite being just a two-year-old startup, Qodo was included in the Gartner Magic Quadrant for AI Code Assistant, an acknowledgment of our capabilities and growing relevance in the software and AI sectors.
• Gartner Cool Vendor
  Qodo was also recognized as a Gartner Cool Vendor,  given to emerging companies that are bringing innovative solutions to their respective industries.
• People’s Choice Award at AWS Unicorn Tank
  At AWS re:Invent, Qodo was the runner up in the AWS Unicorn Tank pitch competition, which showcased promising startups.
• Acceptance into the AWS Gen AI Startup Accelerator
  Qodo was selected for the AWS Generative AI Startup Accelerator program, which provided mentorship and resources to support the company’s work in generative AI.

These accomplishments marked an important year for Qodo, showcasing our steady growth and recognition as a startup making an impact in AI.

Qodo is growing! 

Qodo is growing, and it’s been an exciting year! Starting with just 22 employees, we’re closing the year with a team of 51, nearly doubling in size. To accommodate this growth, we’ve moved into a larger office that reflects the vibrant and dynamic culture of our company. With team members now spanning the US, Europe, and Israel, Qodo is evolving into a global organization, bringing together diverse perspectives to drive innovation and success.

And we’re not stopping here—Qodo is hiring! Join our journey and check out our open positions today.