Building an AI Software Engineering Workspace
For the past few months, I have been working on Navigator.
The objective here is not to create another AI chatbot or another code generator.
The objective here is to develop an AI-powered software engineering workspace that integrates AI into the entire process of developing software.
Navigator is being built to:
• Understand complete codebases rather than individual prompts.
• Debug complex applications.
• Change multiple files at once.
• Navigate through project architectures automatically.
• Recall project context in different sessions.
• Run projects on cloud infrastructure.
• See previews of the changes made.
• Support large codebases.
• Resume unfinished work.
• Create production-ready features.
• Work with project history.
• Allow developers to quickly iterate from idea to deployment.
Rather than jumping from one tool to another, the idea is to encapsulate the whole software engineering process into a single workspace.
Still under development, but each milestone takes me closer to this idea.
It would be great to know which feature will help you use AI workspace effectively in your workflow.
Replies
The decision not to build another chatbot or basic code generator caught my attention.
What is the workspace responsible for beyond code generation—architecture context, task planning, repository navigation, testing, or coordinating changes across multiple files?
For me, the biggest value of an AI engineering tool is maintaining context across a real codebase rather than producing isolated snippets.
I need some input on two points above :
Change multiple files at once: so is it like it will create whole plan first before coding, cause if one component is dependent on another file and that is not being built yet, so how is it going to do that.
Resume unfinished work: going to save context for longer time ? Cause currently in claude code if we start same session after an hour it loads whole context again and that goes towards usage.