We built PorchOps because nobody was going to build it for us.
A generation of solo founders is shipping real software on Lovable, Replit, v0, Bolt, and Cursor, charging real money for it, and running the rest of the business out of five tabs and a shared inbox. The model layer is becoming a utility; the back office that decides whether any of it survives the next quarter still doesn't exist for them. PorchOps is what that back office looks like when the people building it have been the people running it.
We've run the back office before.
We've been ICs, Managers, Directors, VPs, and SVPs at SaaS businesses of various sizes and maturity, across almost every function: engineering, product, project management, customer success, infrastructure and IT, and security and compliance. We've lived the back office. We know what good looks like, and we're codifying that and democratizing it in one platform.
None of us is here for the novelty of putting AI on something. We're here because we spent the last decade and a half watching the same operational work get redone badly in spreadsheets, in Notion docs, in Slack threads that aged into landfill — and the same load-bearing playbooks ride out of every company on the laptops of the people who wrote them. That's the thing PorchOps is meant to fix.
How we work.
The crew runs in drafts by default. Lou, Frankie, Hank, Dale, and Inky pick up the work, do the research, write the response, file the entry, and stop one step short of sending. You read what they drafted and approve it before it goes out. Turning that off, per playbook, once the crew has got it right enough times that the approval click stops feeling like a check, is where this is going.
Every action the crew takes lands in an audit log that reads like a paper trail rather than a vector store. You can see what they read, what they wrote, what they were about to do, and what you said when you stopped them. The stop is per playbook: pause one and it halts at its next step, with your approval queue and audit trail exactly where they were.
Ghost mode is how PorchOps staff reach into your workspace when you ask for help, and it's the only way: read-only access, time-bounded, with a written reason on the record. You see the banner the entire time it runs.
We will never train AI models on customer data — a binding commitment in our privacy policy. The content we send to model providers through their APIs is not used to train their models under their API terms. If you ever can't find it where you expected, that's a bug, and we'd like to hear about it.
Get in touch.
If you want to talk to us, email [email protected]. One of us reads it.



