I’m building Agentic Control Freak (ACF): a local control plane for Codex CLI, Claude Code, Antigravity CLI, and Ollama. With ACF, you can build local projects ranging from Next.js apps to training ML models all via your browser on your local machine.
The agents still do the coding. ACF owns the outer loop:
"task -> plan -> approve -> execute -> verify -> preview -> handoff"
It keeps durable state outside any one agent session: plans, tasks, filesystem diffs, checkpoints, undos, forks, verification runs, previews, memory, skills, and handoff briefs. That means you can switch providers mid-project without starting from zero.
It also does not trust the agent’s final message. The filesystem diff is truth; verification and live preview are owned by the control plane.
The orchestration, workspaces, previews, and state are local: localhost-only, workspace-confined under .workspace/<project>, and not meant for hosting. Provider CLIs may still call their model services. Optional Telegram control lets you approve or steer runs from your phone through a token-gated local worker.
Repo: https://github.com/Antibody/Agentic-Control-Freak
Video intro: https://youtu.be/1fRH-XQrgkY
ML model design and training video: https://youtu.be/KdiiU4RIfFU
P.S. I’d love feedback, especially on whether the app works on Mac or Linux (I tested only on Win11).
I don't get it. Why do people want to use agents to do the coding?
The coding is the fun part.
I'm just speaking for me, as a senior software architect with a decade of experience:
Typing all of that out (for momths and years mind you) is the boring part. For me typing is what comes after thinking. My brain goes into that adhd state where I have to pay enough attention to type the code, but not enough to be engaged. I'm getting so bored while typing code, I can't sit still for 30 seconds. Time gets slow. I get uncomfortable. I keep being distracted. I start thinking about something else and start making stupid mistakes.
I much rather explain my vision, architecture and solution in a detailed text (which helps me think about it in more detai) and then do hours of code review (guess I'm weird for liking that) and "thats wrong, redo it" until I'm happy then type all of that out myself.
There is a lot I hate about "AI", but sparing me from that rage-inducing manual slave labour is not one of those.
If I could hire a human to do the typing I'd do it. But an LLM subscription is so much cheaper than even the worst apprentice.
I've been coding for probably twenty years and I'm still enjoying it.
The problem solving part you're talking about is fun too — it's the social part of development. You get to talk to people and figure stuff out together. What's our common interface, what does the customer want, nay need; security aspects, synchronization problems, etc.
Then you get to work on your craft. You make it nice. You make it tight. You make it lean. And you get to just zone in on your own for a while, and you're in full control. Everytime you build something large and new, you're most likely working with a brand new stack with lots of QoL improvements and your joy of coding rises yet again due to higher DX.
So I like both things.
But the reviewing and correcting other's mistakes? So tedious.
Also in my experience, the thinking and problem solving is the stuff that takes the longest. The actual coding part is only a small part of the development process.
It is indeed all fun and pleasure if you are good at it and love it
I don't know if I love it, but it's a lot of fun. More fun than reviewing the code and fixing all the bullshit in the output. That's literally the worst part of software development.
I usually also review my own code, and when I wrote it myself, it's orders of magnitude faster because I recognize what I've written, too, so I'm minimizing the boring part as well. And being more efficient.
I just don't get it.