Hello Cursor Team,
I wanted to share some feedback from a long-term real-world project where Cursor Pro+ has become an essential part of both research and production operations.
I am the founder of a project called SignalART.
SignalART is not a traditional software product. It is a multi-year investment research and operational framework focused on survivability, governance, replay integrity, diagnostics, and executable portfolio management.
Over the years, the repository has grown into a large ecosystem consisting of:
-
Research frameworks
-
Governance systems
-
Forensic audit tools
-
Operational workflows
-
Backup and recovery infrastructure
-
Website publishing systems
-
Portfolio allocation engines
-
Production monitoring artifacts
The project now contains thousands of files and a significant amount of accumulated institutional knowledge.
What surprised me most is that Cursor gradually stopped feeling like a code editor.
It became an engineering and research partner.
In the early stages, I used Cursor mainly for implementation assistance. Today, I routinely use it for:
-
Architecture reviews
-
Governance audits
-
Forensic investigations
-
Production safety checks
-
Backup validation
-
Documentation generation
-
Repository-wide analysis
-
Long-term maintenance planning
A recent example illustrates this particularly well.
I asked a simple question:
“Is my backup system actually protecting SignalART?”
That question led to a full investigation.
Cursor examined the repository, inspected historical backup archives, compared backup structures with the current production environment, and discovered that the existing backup process was protecting only a legacy subset of the project rather than the current production system.
From there, Cursor helped design a complete preservation architecture, implement the solution, validate the results, generate operational documentation, and establish a three-layer backup system consisting of local storage, cloud storage, and offline SSD storage.
The final result was not simply code generation.
It was a complete operational solution.
What stands out most to me is the importance of context capacity.
As SignalART grew, the challenge was no longer writing code. The challenge became understanding relationships across hundreds of files, historical decisions, governance rules, operational procedures, production artifacts, and years of accumulated research.
With Cursor Pro+, I have repeatedly observed the agent working successfully across roughly 350–400 files of active project context while maintaining coherence throughout an investigation or implementation cycle.
In one recent case, the agent traced relationships across backup systems, governance artifacts, operational workflows, website infrastructure, and production repositories to identify a hidden survivability issue.
The investigation ultimately led to the discovery that the existing backup process was only protecting a partial legacy environment rather than the current production system.
The agent then helped design, implement, validate, and document a complete preservation workflow, resulting in a fully redundant backup system across local storage, cloud storage, and offline archival storage.
This level of repository-wide reasoning would have been extremely difficult to achieve through isolated code generation alone.
For large, long-lived systems, that capability is transformative.
The value is not simply faster coding.
The value is maintaining understanding of an evolving system over time.
SignalART is still under active development and will likely continue evolving for many years before reaching its final form. Cursor has become a meaningful part of that journey.
I wanted to share this feedback because many people evaluate AI tools primarily through coding speed or code completion quality.
My experience has been different.
For large research systems, long-term continuity, repository awareness, operational reasoning, and the ability to carry context across a complex codebase have proven even more valuable.
Thank you for building Cursor and continuing to push the boundaries of what an engineering agent can do.
For anyone interested in seeing the project in practice, SignalART publishes its research, methodology, operational framework, and ongoing results publicly.
Website:
Many of the governance systems, forensic audits, operational workflows, backup architecture, and research infrastructure described in this message were developed with Cursor’s assistance over the course of the project.
SignalART is not a prototype or coding exercise. It is an actively maintained research and operational system that continues to evolve through ongoing research, validation, and real-world operation.
I look forward to seeing how Cursor continues to evolve.
Best regards,
Roku
Founder, SignalART
Independent Researcher