FAQ
Frequently Asked Questions
What is Superpositional?
Superpositional gives engineers and AI coding agents a shared working memory of their codebase. It maps how your code fits together and keeps that understanding current as it changes, so you can plan, build and review with the context of the wider system — not just the files in front of you.
Who is it for?
Superpositional is built for small engineering teams that build with AI. It’s especially useful when one or two people hold most of the system knowledge, agents need repeated explanations, or reviewing changes is becoming a bottleneck.
How does it work?
Connect your GitHub repository through the Superpositional GitHub App. Superpositional indexes the code and its dependencies, then updates that map on every push. Engineers and agents can draw on that context through chat, MCP tools and code reviews.
How is it different from my AI coding assistant?
Your coding assistant helps you write and change code. Superpositional helps it understand the system it’s changing — and provides another check on the result. Keep using your existing agent; give it access to shared codebase context rather than explaining everything again for each task.
Is it just another AI code reviewer?
No. PR review is one part of it. You can also check a plan or specification before coding, explore what a change could affect, ask questions about unfamiliar code, and review a diff before pushing. These workflows draw on the same codebase context.
Can I use it with Cursor or Claude Code?
Yes. Superpositional exposes context and review tools through MCP, a standard that lets compatible AI tools connect to external services. You can use those tools from MCP-compatible clients such as Cursor and Claude Code without replacing your coding workflow.
What can it check before I write code?
Superpositional can review a plan or specification against your codebase, architecture and documented direction. It can also trace dependencies to help you understand what a proposed change might affect — before you invest in the implementation.
Can it follow our team’s architecture and conventions?
Yes. Reviews can use repository-specific rules and documented decisions, including architecture decision records and DIRECTION.md. This gives Superpositional context about how your team intends the system to evolve, not just how the code looks today.
What are signals and concerns?
They’re tracked risks in your codebase, rather than comments on a single change. They help you see what needs attention and how risks are changing over time, with supporting evidence to investigate further.
Do I need to use pull requests to get value?
No. Through MCP and chat, you can ask codebase questions, check plans, explore dependencies and review local diffs before pushing. Automatic PR review is available when pull requests are part of your workflow.
Does Superpositional replace human review or testing?
No. It helps you investigate changes and decide where to focus your attention, but it can miss problems or get things wrong. Keep your tests, CI checks and human judgement in place. A favourable review is not a guarantee that a change is safe to ship.
Does it write fixes or deploy code for me?
Superpositional’s current focus is codebase understanding, risk detection and review — not replacing your coding agent or deployment pipeline. Your engineers and agents make the changes, and your existing CI/CD tools handle deployment.
Which code-hosting platforms do you support?
Superpositional currently supports GitHub. If you use GitLab or Bitbucket, contact us to discuss what you need.
How do I get access?
Superpositional is currently in beta — we’re launching on Tuesday 27 October. Join the waitlist on our website, or email grace@superpositional.io to discuss your team and workflow.