[SUCCESS] Full AI-Ops Achieved: FastAPI Rebuild (Design to AWS Deploy) with Cursor 2.0 + Composer 1

Hi Cursor Team and Community,

I wanted to share a major workflow breakthrough I achieved using the latest Cursor stack.

I successfully completed a full rebuild of a production-grade FastAPI system—from the initial V2 construction plan to the final AWS deployment—entirely via natural-language interaction within Cursor. The human intervention was limited only to necessary HITL approvals.

This was not just “assisted coding”; it was a fully orchestrated, human-governed AI-Ops workflow.

The Breakthrough: Why 2.0/Composer 1 Made it Real

I believe this level of end-to-end autonomy was enabled by the key innovations in the latest release:

  1. Structured Planning (Cursor 2.0): The ability to generate a structured construction plan (V2_CONSTRUCTION_PLAN.md) and maintain context across continuous steps (e.g., code copy, dependency updates, GitOps) dramatically improved multi-step reasoning and continuity.
  2. Agentic Coherence (Composer 1): Composer 1’s agentic context handling maintained deep consistency across complex tasks like refactoring the MeCab-to-Unidic migration, file edits, Git operations, and testing.
  3. Integrated Governance (HITL): The integrated Plan + Chat workflow enabled seamless Human-in-the-Loop (HITL) approvals through natural dialogue, transforming the workflow into a safe, auditable process.

Summary of Achievement

  • Task: FastAPI v2 system rebuild (including GPT-5 model switch and MeCab unidic integration).
  • Execution: 100% completed via natural-language instructions and HITL approvals.
  • Final Output: Automated documentation (V2_CONSTRUCTION_PLAN.md) and successful deployment to the test environment.

Having used Cursor since version 1.x, I can confidently say that 2.0 finally fulfills the long-standing vision — “developing software as naturally as speaking.” Thank you for creating such an exceptional tool and for turning that ideal into reality.

I’d be happy to share any non-confidential technical details or documentation that would be helpful for your internal validation.

— Tomoki Shibahara
AI Systems Architect / Japan @ Independent Research

Focused on AI-driven web service development and human-governed orchestration systems integrating LLMs, FastAPI, and AWS automation.

These are screenshots captured in a Japanese execution environment, sharing part of the actual workflow in progress.