Overture: a local-first control plane built around Symphony

[mefree2098]

I've been building Overture on top of Symphony and wanted to share it here in case it's useful to anyone else working on AI-assisted delivery.

The problem I was trying to solve was pretty simple: I've been using what I call "Prompt Stacking Spec Development" using AI models for a year now, and I wanted one control plane that could start from either a finished plan.md or a rough brief and then carry the work forward. To be clear, I am not a developer in any way, shape, or form. When I saw Symphony I saw the perfect opportunity to make this single platform a reality.

Overture supports both paths. If I already have a solid plan.md, I can paste it in and go straight to execution. If I only have an idea, it walks through prompt workshop, deep research, plan review, Symphony execution, local launch, and deploy.

A few things that have mattered in practice:

• plan.md stays the canonical handoff
• the runtime is local-first
• it can reuse local ChatGPT-backed Codex auth
• Symphony runs, gates, findings, artifacts, and token usage all show up in one dashboard
• launch and deploy are part of the same flow instead of an afterthought

Repo: https://github.com/mefree2098/overture

I'm attaching a couple screenshots below. One shows the intake flow, and the other shows the project dashboard after ingesting a plan.

It's still rough around the edges, but it's already useful for turning either a real plan or a messy brief into something Symphony can execute and track end to end.

Ideally in the end anyone could clone this, deploy it, and build anything from an entire SaaS platform, a website, or an app and deploy without knowing anything about coding.

FYI, this was built using GPT 5.4 in codex, using prompt stacking in ChatGPT chat/deep research for the initial codex deployment plan.

If anyone here is building with Symphony, I'd love feedback on the orchestration model, tracker surface, or what you'd want from a control plane around it.