Sparkles, a San Francisco–based startup building an AI-native workflow platform that enables product, design, marketing, and operations teams to safely make changes to existing production codebases, has raised $3.5 million in seed funding as it emerges from early-stage development and joins the Y Combinator ecosystem.

The round includes backing from Y Combinator, along with participation from Moonfire Ventures and a group of angel investors from leading AI and infrastructure companies, including OpenAI, Anthropic, Andreessen Horowitz (a16z), General Catalyst, and Cloudflare.

Founded by Dan, a University College London dropout and former collaborator with the former CTO of Monzo at Iterate.com, Sparkles is building what it describes as a “Lovable for large existing codebases,” aiming to bridge the gap between non-technical teams and production engineering systems. The platform is designed to allow users to generate changes to live software through GitHub-native workflows, where outputs are converted into reviewable pull requests with built-in safeguards and guardrails.

Sparkles positions itself in a rapidly expanding category of AI development tools that focus not on writing new code from scratch, but on modifying, maintaining, and operating large production systems. The company argues that while AI coding assistants have become widely adopted, most organizations still rely heavily on outsourced engineering or bottlenecked internal teams to ship even small changes to production environments.

Its core product enables non-engineering stakeholders such as product managers, designers, marketers, and operations staff to propose and execute safe changes to production systems without requiring deep technical expertise. These changes are processed through controlled workflows that generate structured outputs, automated checks, and GitHub pull requests that engineers can review and merge.

The company’s approach is built around the idea of “safe autonomy,” where AI systems assist in production changes but remain constrained by reviewable, auditable workflows. Sparkles emphasizes that its platform is not intended to bypass engineers, but to reduce the friction between idea generation and implementation in large, complex codebases.

The $3.5 million seed funding will be used to expand engineering development, improve its AI workflow engine, and grow integrations with major development tools and cloud infrastructure providers. The company also plans to invest in strengthening its safety and reliability systems, particularly around permissioning, code validation, and deployment safeguards.

Sparkles is part of a broader wave of startups emerging from the Y Combinator Winter 2026 batch that are focused on AI-native infrastructure for software development and enterprise productivity. Its positioning reflects increasing demand for tools that not only accelerate coding, but also democratize access to production systems across non-technical roles.

With support from Y Combinator and a network of high-profile investors from leading AI labs and venture firms, Sparkles is aiming to establish itself as a foundational layer in the next generation of software development workflows, where AI agents act as intermediaries between business intent and production code execution.

Related Articles