Blitzy launches sandbox offering free autonomous software development for enterprise codebases
Enterprise autonomous software development platform Blitzy today announced the launch of the Blitzy Sandbox, a trial environment that allows organizations to evaluate the platform on their own software estates at no cost. The offering enables eligible enterprises to ingest up to 1 million lines of code and generate up to 25,000 lines of forward-engineered, end-to-end tested code as pull requests.
Blitzy is built specifically for complex enterprise applications that span multiple repositories, legacy architectures, and undocumented systems, helping teams accelerate software delivery by 5x or more.
Understanding Enterprise Codebases at Scale
Unlike standard AI coding assistants that focus on single-file suggestions, Blitzy builds a system-level knowledge graph of an entire software estate. By ingesting millions of lines of code, the platform constructs a semantic representation of business logic, dependencies, architecture, and data flows. The platform remains code-agnostic and language-agnostic, storing representations of system intent rather than the underlying raw code.
“Where our platform shines is large enterprise estates” Blitzy’s Dan Ochs, head of technical customer delivery, told SD Times. “So think millions of lines of code, and what we do is we have a semantic understanding of that code, and the way we do that is by ingesting the code, and we have a robust knowledge base that we create from that code, which is a representation of what the code does semantically or from a business perspective. So, from that large-scale understanding, we can then work with customers, or customers can work with the platform to then execute tasks against that large code based, and so you could do, for instance, a COBOL to Java migration, and we have customers that have done that here, saved 4300 hours of engineering time for a byte to byte identical to the original mainframe, and that’s 500,000 output lines of Java.”
To power code generation and transformation, Blitzy employs a multi-model orchestration framework using foundation models from providers including Google, Anthropic, and OpenAI. Thousands of autonomous agents are spun up in parallel to execute complex development plans, validate logic, and run end-to-end testing. The platform also is language-agnostic, he said.
Refactoring, Migrations, and Greenfield Development
While Blitzy frequently engages with enterprise estates burdened by decades of undocumented legacy code, its platform supports a wide spectrum of software engineering tasks, which Ochs categorized as “forward engineering”:
- Legacy Modernization: Performing multi-repository refactoring or mainframe-to-modern language migrations (such as converting legacy COBOL to Java with byte-for-byte identical behavior).
- Cross-Platform Conversions: Translating code bases between platforms—such as converting native iOS mobile code to Electron to expand market reach onto Android and desktop environments.
- Greenfield & Feature Development: Building new services and applications from scratch, supported by direct integrations with design tools like Figma to transform product designs into opinionated architecture and code.
- Testing & Documentation: Generating missing architectural documentation, mapping system dependencies, and building test suites to reach high coverage targets.
Proactive Insights: Vulnerability Detection and Remediation
As part of the platform, Ochs said Blitzy incorporates Proactive Insights, an automated security feature that scans ingested code for vulnerabilities. According to the company’s blog post announcing the Sandbox, the platform uncovers proprietary zero-day security risks and reconciles findings from existing scanners, such as Snyk, Mend, or GitHub Advanced Security.
Going beyond detection, Blitzy analyzes execution paths to determine whether a vulnerability is reachable and exposed to external endpoints. It evaluates the architectural impact of a potential fix and automatically generates remediation pull requests, enabling security and development teams to resolve flaws in a single workflow.
Enterprise Trust and Production Integration
Blitzy operates as a complementary layer to existing development pipelines. To further differentiate the platform from AI coding assistants, Ochs explained: “The output from our platform will get as far as we can, as close to production as we can. But we create what’s called a project guide and a pull request. And in that project guide, it explains what the engineering team needs to do in order to go from pull request into a live production environment. And that’s where the engineering team will come back in, along with the AI coding tools, to do that last mile to get into production. So I would say we’re complementary to those tools.”
To safeguard enterprise IP, customer code is never used to train foundation models, and execution takes place in sandboxed environments built to SOC 2 Type II and ISO 27001 compliance standards.
Organizations can access the Blitzy Sandbox today to evaluate the platform across legacy refactoring, security remediation, and new feature development on their own codebases.
The post Blitzy launches sandbox offering free autonomous software development for enterprise codebases appeared first on SD Times.
Tech Developers
No comments