Resume Studio

A structured resume editor with reviewable AI edits through MCP, shared schemas and transactional revisions.

2026TypeScriptNext.jsReactZodSQLite / PostgreSQLSupabase AuthCloudflare WorkersHyperdriveBrowser RunMCPPlaywright

Problem

Many people now use AI to update their resumes or tailor them to a job. Others, worse, don’t use it at all. Either way, updating your resume for each job is tiresome. Even with AI, you read through what it says, then make every change yourself. It’s needlessly complex.

What I built

An MCP app with a frontend. You plug in your own AI and it works seamlessly with the editor, giving you a resume updated for the job. You can review each change or let the AI apply it, and every change is recorded and can be restored.

Underneath is a structured resume editor with an API and MCP interface. Shared schemas and deterministic operations span all three entry points. Changes use transactional edits and immutable revisions, with scoped proposals and protection against stale edits.

Architecture

EditorYou, in the browserLive preview while you edit AI hostYour AI assistantClaude Code, Codex, Cursor Entry pointAPIChecks origin and input Entry pointMCP server27 tools on the same operations ValidateShared schemas ProposeBefore and after ReviewYou, or the AI CommitOne transaction SnapshotRestorable Shared coreOne definition of every edit OutputPreview and PDFThe same pages in both OutputRendererOne set of templatesand pagination rules StorageStructured dataResumes, versions,proposals, facts HTTP MCP EditorYouIn the browser AI hostYour AIClaude Code, Codex Entry pointAPIChecks every call Entry pointMCP server27 tools Shared coreOne definition of every edit ValidateShared schemas ProposeBefore and after ReviewYou, or the AI CommitOne transaction SnapshotRestorable StorageStructured dataResumes, versions, proposals, facts OutputRendererOne set of templates and pagination rules OutputPreview and PDFThe same pages in both HTTP MCP

Key features

  • Bring your own AI. Connect Claude Code, Codex or Cursor over MCP.
  • Tailored copies. Clone your master resume for each job and keep the original intact.
  • Review or auto-apply. See before, after and the reason for each change, then accept, reject or let the AI apply it.
  • Live preview. Watch the page update as edits land.
  • Five templates. Switch the design without touching your content.
  • Page fit check. See how a change fits on the page before it goes in.
  • Version history. Every edit is saved and can be restored.
  • PDF export. The download matches the preview.

Quality & testing

The project uses strict TypeScript, ESLint, CI and SQL migrations. Node tests and smoke scripts validate the implementation; Playwright is part of the toolchain.

What’s next

Today, Resume Studio works with an AI you connect over MCP. Next, it becomes a full standalone app.

  • Built-in AI chat. Ask for changes inside the editor, without connecting another app.
  • Backend LLM integration. The app runs the AI itself, so there is nothing to connect first.