How I Priced My First Claude Code Project (And What I'd Charge Today)

My first paid Claude Code project was a $500 flat fee for a real-estate agent who wanted a single-page website with a property gallery and a contact form. The whole build, including the back-and-forth on copy, took me about six hours over two evenings. I sent the invoice, she paid same day, and I felt like a genius. I was not a genius. I had just sold a $3,500 outcome for $500 because I priced what I knew - the time - instead of what she actually got, which was a real online storefront for her business and three weeks faster to launch than the local web shop had quoted her.

I'm telling that story up front because everyone undercharges on their first Claude Code project. Every single person I've watched go through this. The reason is structural - when you're new, you can only see the inputs (your time, your effort, your nerves) and you can't yet see the outputs (the client's relief, the revenue they unlock, the months of decision they avoid). Until you can see the outputs clearly, you'll keep pricing the inputs. That's the whole trap.

The point of this piece isn't to make you feel bad about your first invoice. It's to compress the lesson so your second invoice is closer to the right number. I'll walk through what I actually charged on a handful of early projects, what the same scope would cost today, and the three-shape pricing system that runs my agency now.

Pricing freelance Claude Code work comes down to three shapes - and you almost never want the first one. Hourly is the default trap because it feels safe and familiar, but every quarter the model gets faster and your hourly invoice shrinks for the same outcome. Project pricing fixes a number to a deliverable, which is where most of my work lives now. Retainers turn a one-time build into ongoing infrastructure for the client, which is where the actual business gets built.

The instinct as a beginner is to default to hourly because you've seen agencies and dev shops do it. Don't copy the dev shop. A dev shop bills hourly because their unit cost is human hours and those hours move slowly. Your unit cost is a Claude Code session that moves at a different speed entirely. Selling that as an hourly rate is selling your ceiling on day one. The cleaner play is to define a scope, name a flat fee, and let the model speed compound in your column.

The shape to pick depends on the work. A landing page, a Stripe checkout, a one-off internal tool - fixed-scope project. An automation that runs every day forever and might need tweaks as the client's process evolves - monthly retainer. An identical productized deliverable you can sell to dozens of clients with light customization - per-deliverable pricing, usually $500 to $3,000. Hourly stays in your toolbox for paid discovery (a $300 audit call that gets credited if they engage) and nothing else.

Let me get concrete. Below are three scopes I've actually shipped, what I originally charged, and what the same scope goes out the door at today. These aren't theoretical numbers - they're the prices that have actually cleared on real invoices.

Scope one: a single-page small business website with a hero, three content sections, a gallery, and a contact form that emails the owner. My first invoice for this was $500. Today the same scope is a $1,800 productized package with a two-week delivery window and a one-month tweak period bundled in. The reason the price tripled isn't that the build got harder - it got easier. The price tripled because I learned to anchor on what the client gets (a live web presence by next Friday) instead of what I do (six hours of building).

Scope two: a Gmail-to-CRM automation that reads inbound leads from a form, scores them, and pushes the qualified ones into a sales tool with a daily digest emailed to the founder. First version of this I built for $1,200 as a one-shot project. Today this is a $2,400-a-month retainer that includes maintenance, model upgrades, and one new feature request per quarter. Same code under the hood. Completely different framing on the invoice, because the client isn't buying a script - they're buying ongoing lead routing as infrastructure.

Scope three: a full agency engagement for a 12-person services business - three connected automations (lead routing, client onboarding, weekly reporting), an internal dashboard, plus quarterly upgrades and a Slack channel for support. This is where you land in the $15K to $25K range, paid as either a six-month fixed engagement or as a recurring $4K-a-month retainer with a $5K kickoff fee. The work is bigger but the multiplier comes from the relationship: you're now the AI partner for a whole business, not the freelancer for a single deliverable.

Value-based pricing - charging a percentage of the outcome rather than a fee for the build - sounds great in books and is genuinely hard in practice. The reason it's hard is that it only works when you can clearly measure what the client gained, and most early projects don't have that data clean enough to anchor on. So the realistic answer is: don't lead with value-based pricing on your first ten projects. Lead with fixed scope or retainer, and let value-based emerge as a third tier once you have a track record.

The signal that you're ready to quote a value-based number is when the client can tell you, on the discovery call, exactly what the current state is costing them. "We lose three hours a day on this handoff." "Our conversion rate is 1.2% and every point is worth $40K." "We're paying a part-time bookkeeper $1,500 a month for work that should be automated." Once you hear a number like that, you have a real anchor. A quote of 20-30% of the first year's saving is generally easy to defend and generous enough that the client doesn't blink.

Until you hear that number, package the work. Three tiers - good, better, best - at flat prices, with the middle tier being the one most clients pick. That packaging move alone will move your average invoice up about 60% over leading with a single number, because clients land naturally on the middle option when there's a third one to anchor against. It also makes the sales conversation about which version they want instead of whether the price is reasonable, which is the only conversation you actually want to be having.

Packaging is the single biggest move you can make to raise prices without raising friction. Instead of quoting one number, you present three versions of the same offer at three price points. The math the buyer is doing changes completely - they're no longer asking "is this worth it?" they're asking "which of these three is the right fit for me?" That second question closes far more often than the first.

Price the tiers so the middle is the obvious value - the starter should feel a little thin, the pro a little ambitious, and the standard should feel like the smart move. Most clients self-select into the middle without you steering them, which is exactly what you want. The ones who pick pro are the relationships that compound for years. The ones who pick starter were never going to be your best client anyway, and now you've at least gotten paid to find out.

If I could send one message back to the version of me about to send that $500 invoice, it would be: triple it, and put it in a one-page scope-of-work before you start. The price would have cleared at $1,500 with zero negotiation because the client was thrilled at $500 - she was comparing me to the $4,000 local web shop, not to free. I left $1,000 on the table because I priced relative to my own anxiety instead of relative to the alternative she was actually weighing.

The other thing I'd tell my younger self: stop treating the first project like a referendum on whether you're "worth" professional rates. The first project is just the first project. You'll learn more about pricing in the second engagement than the first, and more in the third than the second, because each one teaches you what the client actually valued and what they wouldn't have paid an extra dollar for. That feedback loop is the thing - not getting the first number perfectly right.

And the last thing: if you're sitting on the first invoice and the number feels too low, it is. Raise it before you send. Almost nobody pushes back on a price that's already inside their budget envelope, and the price that feels uncomfortable to you is usually still well inside theirs. That single adjustment - sending a number that scares you a little - is what moves you from hobbyist rates to real freelance rates faster than anything else I know. I'm still inside the club every day at $9 a month learning from people who price better than me, and that compounding is the actual answer to where the rates come from.