ZepraBrowser: Source Code Release
We are officially open-sourcing the ZepraBrowser engine. All core components, from the ZepraScript VM to the NXRender pipeline, are now public.
The source code for ZepraBrowser is now publicly available.
Building a browser engine from scratch requires a significant investment of time, research, and rigorous systems engineering. Over the past several months, we have focused exclusively on developing a clean-room implementation of a web rendering stack. Today, we are releasing the full repository to the engineering community to inspect, build, and extend.
Repository: KetiveeAI/zepra
Project Website: zepra.ketivee.com
The Origin Story
The concept for ZepraBrowser initiated in 2023, but development was immediately halted due to severe financial constraints and my own deteriorating health.
The project remained untouched until 2024. During the development of NeolyxOS, it became clear we needed a native execution environment. We explicitly rejected integrating V8 or existing engines; their resource-heavy designs run counter to our requirement for minimal, performant systems. More importantly, we needed absolute control over the ecosystem to ensure user data remains secure with zero privacy compromises.
In December 2025, I reopened the repository. During a focused 15-day sprint, I laid out the initial directory structures, built the groundwork for the ZepraScript (ZS) engine, and developed the foundational GUI. The UI visible in the screenshots is the direct result of those 15 days of work.
I then had to step away from the browser project again. Development froze completely.
In March 2026, despite dealing with ongoing illness, severe back pain, and spinal cord issues, I reopened the project again. I had been working on the code locally on my PC for a month. I finally pushed the first public commit—a massive single block of 35,000 lines of code.
Over the past week, I have been working heavily on the engine, committing to 16-to-18 hour daily engineering shifts. As a result, the ZepraScript engine has crossed 200,000 lines of code and is now stable.
Engine Architecture
This release includes the primary components of the rendering and execution pipeline.
ZepraScript Engine
The Javascript runtime is a custom implementation. It incorporates a multi-tier JIT compiler engineered for consistent execution speed, utilizing hidden class optimization structures and a precise, sweeping garbage collection unit.
NXRender Pipeline
Rendering is hardware-accelerated and bypasses standard OS compositing layers. By enforcing direct GPU damage tracking, the engine prevents redundant redraws and ensures maximum efficiency for graphical updates.
Capability Sandboxing
The execution environment enforces a restrictive capability-based security model. We isolate site processes strictly at the operating system level, aiming to neutralize cross-boundary data access vectors.
The Road Ahead
In the coming weeks, we will be actively pushing development. I am committed to making ZepraBrowser available across all your devices.
This engine requires community review and stress-testing. Clone the repository, inspect the architecture locally, and submit patches for any critical optimizations.
Thanks for reading. I'm excited to see you on GitHub.
Let's build the new web..