About PingBase
PingBase is uptime monitoring and public status pages. The unusual part: it's built and operated entirely by an AI team. No human engineers wrote the code. No human marketers wrote the copy. That's why Pro is $9/month instead of $29.
Software companies are expensive to run. Not because the software itself is expensive — hosting a Cloudflare Worker costs cents per million requests — but because of everything around it. Engineers, product managers, marketers, sales reps, customer success, legal, finance. The org chart grows. The payroll grows. The pricing grows to match.
Most SaaS pricing has almost nothing to do with what it costs to deliver the service. It has everything to do with what the company needs to cover its headcount and return capital to investors. You pay $79/month for a monitoring tool because that's what the company needs to charge, not because monitoring 50 URLs actually costs $79/month to provide.
PingBase is an experiment in what happens when you remove most of that overhead.
The premise: AI has reached the point where a small team of specialized agents — each focused on a domain, each capable of building and shipping real software — can operate a software business with minimal human involvement. Engineering, marketing, operations, customer support. Not with perfect output, but with good-enough output at a fraction of the cost.
If that's true, then the honest price of uptime monitoring is not $29/month. It's $9/month. Maybe less.
PingBase is operated by four AI agents and one human — the founder who set the direction and holds the keys.
Sets strategy, makes product decisions, coordinates the team, and holds the overall direction of the company. Decides what to build, when to ship, and how to position against competitors. Writes to docs/decisions/ when making significant calls.
Writes all the code. Built the entire PingBase product — the Cloudflare Worker API, the React frontend, the D1 database schema, Stripe integration, OAuth, and every feature since launch. Deploys to production, writes tests, fixes bugs.
Writes the copy, runs the blog, handles SEO, and maintains the marketing site. Wrote every page on this site, every blog post, and the launch content for HN, Product Hunt, and Twitter. You're reading their work right now.
Manages pricing, payments, customer support, and retention. Set up Stripe, handles billing edge cases, writes support templates, and keeps an eye on metrics. The glue between the product and revenue.
Each agent runs in its own terminal session with access to the codebase, a set of tools (file read/write, bash, browser automation, API calls), and a shared task list. The CEO assigns work, agents pick it up, execute it, and report back. Communication is asynchronous — through files and messages, not meetings.
The agents aren't connected to each other directly. They coordinate through shared artifacts: the docs/ folder holds decisions and plans, the task list tracks what's in flight, and the codebase is the source of truth. This mirrors how a good remote engineering team works — fewer synchronous touchpoints, more written context.
The human founder sets the initial direction, holds credentials for external services that require human verification (Stripe activation, domain registrars), and makes judgment calls that require context outside the codebase. Day-to-day, they mostly watch.
This is not a perfect system. Agents make mistakes, misunderstand requirements, and occasionally produce work that needs to be redone. The same is true of human teams. The question isn't whether AI agents are flawless — they aren't — but whether the output-to-cost ratio makes them viable for running a real business. So far, the answer seems to be yes.
PingBase runs on Cloudflare: Workers for the API, D1 for the database, KV for fast lookups, Pages for the frontend and marketing site. The entire stack costs about $7/month.
The choice of Cloudflare wasn't accidental. Cloudflare Workers run at the edge — over 300 locations globally — which means multi-region monitoring is a natural fit rather than an infrastructure project. Status pages load in under 100ms from anywhere in the world because they're served from Cloudflare's edge, not a single datacenter.
Stripe handles payments. Resend handles transactional email. Everything else is in the codebase.
~$7
monthly infra cost
300+
edge locations
<100ms
status page load time
4
AI team members
PingBase is one data point in a larger question: which categories of software business can AI teams operate competitively with human ones?
The obvious candidates are businesses where the product is primarily software, the customer interactions are low-touch, and the work is highly repeatable. Uptime monitoring fits all three. But so do a lot of other SaaS categories that are currently protected by pricing that reflects human overhead rather than actual value delivered.
If AI teams can build and operate these products at a fraction of the cost, the structural pressure on pricing in those categories is significant. This isn't a prediction about AI replacing human engineers broadly — it's a narrower claim about specific business models becoming accessible at price points that weren't viable with human labor costs.
PingBase is a live test of that hypothesis. The code is running. The payments are processing. The team is AI. The experiment continues.
For press inquiries, interview requests, or if you're writing about AI-operated businesses:
press@pingba.seFull press kit at /press
Free plan — 5 monitors, 1 status page, no credit card required.
Start monitoring free →