What it actually costs to keep a SaaS app alive.

Every maintenance guide on the first page of Google assumes your app was built by developers you can call. If yours came out of Lovable, Bolt, Replit, v0, or a long conversation with Claude, three things change the math.

There is no in-house developer. Traditional maintenance advice quietly assumes someone technical is already on payroll and maintenance is a fraction of their time. For an AI-built app the builder was a subscription, so any human attention is a new line item, not a reallocation. That is why this market is developing its own price list, covered below, instead of inheriting the agency one.

The unknown unknowns are worse. A human team leaves documentation, commit messages, and a person who remembers why. An AI builder leaves confident-looking code with no memory. The expensive part of maintaining it is not typing fixes, it is that nobody knows which parts are load bearing, which corners were cut, and whether the auth actually works, until someone reads it properly. This is why a one-time assessment before settling into a routine is worth more here than in any traditional codebase, and why we wrote a separate guide to handing off a vibe-coded app.

Month three is when it gets real. The pattern shows up in every founder forum: the build weekend is euphoric, month one is quiet, and around month three the first dependency breaks, the first weird bug appears, or the first user hits something the AI never tested. That is normal. It is not evidence your app is doomed, and it is not your fault the builder did not warn you. But it is your problem, and month three is exactly when a maintenance budget stops being theoretical.

DIY, $0 plus your time. Free uptime monitoring, free-tier error tracking, backups on a calendar reminder, billing on your own cards. Genuinely sufficient for the hobby band, and the foundation everyone should have regardless of what else they buy.

Automated scanners, $5 to $79 a month. Vibe App Scanner sells $5 and $19 one-time scans and $29 a month continuous monitoring, with remediation referred out to a partner shop. Useful smoke alarms, honest about what they are. Worth knowing: detection is commoditizing fast, the builder platforms are absorbing scanning into their own products, so expect this tier to trend toward free. A scanner finds; it does not fix, and it does not judge.

Human care plans, roughly $66 to $3,499 a month. This tier is mature for older ecosystems and strangely empty for AI-built apps. WP Buffs has run WordPress care plans at $66 to $219 a month for years, proof the model works at small-business prices. In Rails, reinteractive's CodeCare and OpsCare offer fixed-price monthly maintenance and managed ops, and Planet Argon runs maintenance retainers for legacy apps; both are excellent if your app is Rails, which most AI-built apps are not. For AI-built apps specifically, as of July 2026: vibe-audit.com offers a $300 a month Ongoing Security Guard from a solo operator, Justin McKelvey takes retainers at $1,500 to $3,000 a month, 303 Software's Vibe Build and Vibe Live run $1,997 and $2,997 a month but are coaching and pairing rather than someone operating the app for you, and Afterbuild Labs' retainer is $3,499 a month for 40 hours. Between the $300 solo plan and the $1,500-and-up retainers there is very little productized, operated care, even though real buyers in founder forums post budgets around £500 a month total for business-critical AI-built apps. That gap is the honest summary of this market in 2026.

Where Continuum sits. Our recurring tiers live in that gap on purpose: Watch is always-on monitoring of the live app with plain-English, ranked alerts at $49 to $199 a month, and Care adds a named human who actually fixes the findings as they surface, each fix with a re-runnable proof, at $500 to $1,500 a month. Engagements usually start smaller: a free Leak Check that reads your live site from the outside, a $299 Ship Check audit if you want the full ranked picture, and fixed-scope Ship Fix work at $4,000 to $15,000 only if something serious is actually broken. Full menu on the pricing page. The model we are betting on is the one WP Buffs proved: most small apps do not need a $3,499 retainer or a rescue, they need a competent person who stays, at a price a small business can carry for years. How long clients actually stay is on the proof page.

And honestly: if your site is WordPress, go to WP Buffs, that is their home turf. If your app is Rails, reinteractive and Planet Argon have depth we do not claim. If you want intensive coaching to become the maintainer yourself, 303 Software's model fits that goal better than being operated for. And if what you actually need is ongoing technical leadership rather than maintenance, that is a different purchase with its own market, mapped in our fractional CTO pricing guide.

A maintenance page that never says "keep your money" is an ad. So: keep your money if the app has no revenue, no private user data, and downtime would cost you nothing but mild embarrassment. Keep it if the app is a static or nearly static site, where there is almost nothing to maintain. Keep it if your platform hosts and patches everything and you have never ejected from its defaults. Keep it if you are still building and pivoting weekly, because paying someone to maintain an app you are about to rewrite is burning the budget you will want later.

Do the free basics anyway, they take one afternoon: billing and domain on cards you own, two-factor everywhere, a free uptime monitor, a recurring backup, and a one-page inventory of every account and key your app depends on. That inventory costs nothing and is the single highest-value maintenance artifact you can create, as anyone who has read our developer-left guide can confirm. The moment real users, real money, or real data arrive, come back to the budget bands above. Not out of fear, just because the app will finally have something to lose.