How much does custom web software cost in 2026? Honest ranges from a senior engineer

Custom web software in 2026 costs anywhere from $2,000 for a small automation to $150,000 or more for a full platform. Most projects I quote for US small and mid-size businesses land between $5,000 and $50,000: internal tools and dashboards typically run $8,000 to $25,000, client portals $15,000 to $50,000, and SaaS MVPs $25,000 to $75,000. Agencies quote 3 to 5 times more than independent senior engineers for the same scope. The rest of this article breaks down where your project lands and why.

I run Troya Labs, a one-engineer studio, and I have been building and maintaining custom web software since 2015. The ranges below are the ones I actually use when I quote, matched against what agencies and offshore teams charge for comparable work. No “it depends, book a call.” You will still need a conversation for an exact number, but you should be able to budget before you ever talk to anyone.

Why is it so hard to get a straight number?

Two reasons, and only one of them is legitimate. The legitimate one: “custom software” covers everything from a 40-hour integration script to a two-year platform, so a single number would be meaningless. The other reason is sales strategy. Publishing prices filters out leads before a salesperson gets to work on them, so most shops keep pricing behind a call. I would rather you arrive knowing the ranges. It makes the conversation shorter and the fit better.

What actually determines the price?

Five factors move the number more than anything else:

1. Scope, measured in screens and workflows. A tool with 4 screens and one job costs a fraction of a system with 30 screens and role hierarchies. When I estimate, I count screens, forms and states before anything else.
2. Integrations. Every external system your software talks to (CRM, payment processor, accounting, shipping APIs) adds days of work and most of the long-term maintenance surface.
3. Roles and permissions. “Everyone sees everything” is cheap. Four user types with different views, approvals and audit trails can double a build.
4. Data migration. Starting fresh is easy. Importing eight years of inconsistent spreadsheet history is a project of its own, and it is the item buyers forget most often.
5. Who builds it. The same scope has three very different price tags depending on whether an agency, an independent senior engineer or an offshore team builds it. More on that below.

How much does each type of project cost?

These are 2026 ranges for work done by an independent senior engineer in the US market. Multiply by 3 to 5 for a traditional agency.

Project type	Typical range	Timeline	Example
Automation or integration	$2,000–$8,000	1–3 weeks	Sync your CRM with your billing system
Internal tool or dashboard	$8,000–$25,000	3–8 weeks	Replace the spreadsheet your ops team fights over
Client portal	$15,000–$50,000	6–12 weeks	Clients log in, upload files, track status, pay
SaaS MVP	$25,000–$75,000	2–4 months	The first sellable version of a software product
Full platform	$75,000–$150,000+	4–9 months	Multi-role system running an entire operation

A note on the low end: the $2,000 to $8,000 automation tier is the most underrated purchase in this table. One well-placed integration that saves an employee 5 hours a week pays for itself in months, and it lets you evaluate how a builder works before you trust them with something big.

Why do agencies quote 3 to 5 times more?

Cost structure, mostly. When an agency quotes $90,000 for a portal I would quote at $25,000, the difference is rarely the code. You are paying for a project manager, an account manager, a designer, one or two mid-level developers, office overhead and margin on each layer. Some projects genuinely need that structure: if you require a 6-person team working in parallel, an agency is the right buy.

For most small and mid-size business projects, the entire scope fits in one senior engineer’s head. Then the layers add cost and communication hops without adding output. When you work with me, the person who scoped your project is the person writing the code and answering your emails. That is the structural reason my quotes come in where they do; the same logic applies to any experienced independent engineer.

Offshore teams sit on the other side: $25 to $50 per hour looks like a third of the price, and sometimes it works out. The failure mode I get hired to fix is a codebase built by rotating developers with no single owner, where the rework costs more than the original build. If you go offshore, buy a specific team with names and a track record, never a rate card.

What does custom software cost after launch?

Plan for 10 to 20% of the build cost per year. A $20,000 tool should carry a $2,000 to $4,000 yearly budget for updates, dependency upgrades and small improvements. Some years you will spend less. The platform I have maintained since 2015 has had years where maintenance was a few hundred dollars of my time.

Hosting is the smaller line: most business applications run comfortably on $20 to $100 per month of cloud infrastructure. If someone quotes you $500 a month of hosting for an internal tool with 30 users, ask exactly what it pays for.

The honest framing: off-the-shelf SaaS spreads cost over a subscription forever, custom software front-loads it. Custom starts winning when the subscription math breaks, which I cover in a separate article on custom vs off-the-shelf.

What makes the price grow mid-project?

Almost always one of these:

• Requirements discovered late. “Oh, and it also needs to handle returns” in week six costs multiples of the same sentence in week zero. The fix is boring: spend real time on scope before code.
• The integration nobody checked. An API that turns out to be undocumented, rate-limited or missing the exact data you need. I probe the risky integrations during scoping, before I commit to a number.
• Design churn. Rebuilding the same screen three ways because stakeholders disagree. Decide who has final say before the project starts.

A fixed-scope, fixed-price first milestone protects you from all three. If a builder will only work open-ended hourly on a vague scope, that risk is priced in, and you are the one carrying it.

How to budget without getting burned

The approach I recommend to people who call me, whether they hire me or not:

1. Write down the one workflow that hurts most. Not the ten-feature vision, the single thing that costs you hours or clients today.
2. Get it scoped as a standalone first phase. It should land in the $5,000 to $20,000 range for most businesses.
3. Ship it, use it for a month, and let real usage decide phase two. Half the features from the original vision usually die here, and that is money saved.
4. Hold back 15% of your total budget as contingency you do not mention to anyone.

If you want a number for your specific case before committing to anything, the free audit at the bottom of this page is the fastest way to get one: you describe the workflow, I reply with what I would build, what I would skip and a realistic range.

Quick answers

Can I get a working app for $5,000?

Yes, if the scope is one workflow: an automation, an integration or a thin internal tool. No, if it involves multiple user roles, payments and a mobile app. Anyone who promises a marketplace for $5,000 is planning to disappoint you slowly.

Should I pay hourly or fixed price?

For a defined scope, fixed price puts the estimation risk on the builder, where it belongs. Hourly makes sense for ongoing maintenance or genuinely exploratory work. Be suspicious of open-ended hourly on a project with a clear goal.

How long does a custom build take?

From the table above: weeks for automations and internal tools, 2 to 4 months for an MVP, 4 to 9 months for a platform. If a timeline sounds too good, the missing time usually reappears as bugs after launch.

Is no-code a cheaper alternative?

For validating an idea, yes, and I say that as someone who earns nothing from no-code. The costs appear later: per-seat pricing, integration limits and workflows the tool cannot express. Several of my projects are rebuilds of no-code systems a business outgrew, so budget for that transition if you start there.