I live in AI coding tools. Most days at least one agent is open, doing real work with me. And every time my machine restarted, or the conversation got long and compacted, the tool forgot. All of it. I was back to explaining my project from scratch, like we had never met.
One day I rebooted mid-task, opened a different project, and wanted to reuse something I had figured out weeks earlier. It was gone, buried in a chat log I would never reopen. I am a human. I remember things. Why can’t the tool I code with?
The constraint that became the idea
Here is the honest part. I didn’t have money. No budget for servers or cloud bills that grow every time someone uses the thing. But I had a Mac, and Macs are quietly powerful machines with a lot of horsepower going unused. So Oplyr runs on your own machine. Local, free to run, private by default. Not because it was trendy. Because it was the only way I could afford to make something good.
Stay the one who decides
AI is going to write a lot of our code, and I am fine with that. What I am not fine with is code becoming a black box that breaks in production at the worst moment. Not by reading every line. Nobody has time for that. But you should always be able to see what is happening: a plain summary of what a file does, and every change shown to you before it touches disk, so you keep the good ones and send the rest back.
AI writes the code. You stay the one who decides.
Memory, not context
Most tools give you context: a flat file of your recent chat that gets squeezed until it is useless. I wanted memory. The kind that keeps what matters, lets the rest fade, and is still there weeks later. It should be shared by every tool you use, so moving between them never means starting over. And it should not start empty. Oplyr reads what Claude Code or Codex already know about your work and offers to bring it in on day one.
Context is a file. Memory is a brain.
A room of agents, and you conduct
Once your agents share a memory, something fun becomes possible. Different agents lean toward different answers, so why pick one? Put them in the same room, let them compare and disagree, then you make the call. One drafts, another flags what is wrong. Give them one brain and they stop being tools and start being a crew. Nothing runs off on its own. And instead of one big plan that burns fast, two smaller ones you already have can share the same Brain.
Bring your own. Nothing locked.
A lot of tools charge you monthly and decide which AI you may use. You bring the logins you already pay for, to the vendors you already trust, and use them here. I am not standing at the door asking you to pay me for the AI. You bring the engines.
One place for everything
Stop making developers switch. A real terminal, not a cramped pane. Your codebase as a tree or a living graph, with a summary for every file. Your docs, rendered somewhere that is not painful to read. Talk to it or type, whichever the moment wants. The core loop comes first. Once it is dependable, the next thing developers reach for belongs here too.
Why this exists
I built this because I needed it. The tools I admire most came from someone scratching their own itch, and it turning out a lot of people had the same itch. I am not promising Oplyr is that. It solved a real problem for me, and I hope it solves it for you. It is early software, and things change week to week. Use it because it is interesting, not because it is finished.
~DukeOfDune, building Oplyr
