Pricing a Dev Tool When Your Competitor Is Free
QueryBear competes, in a sense, with free. There are open-source MCP servers for Postgres. You can wire up a read-only role and a proxy yourself in an afternoon. I've written posts on this very blog showing people how to do exactly that. So a reasonable person asks: why would anyone pay you for something they can assemble for free?
It's a fair question, and "ours is better" is a bad answer. Here's how I actually think about pricing against free, because I had to get past the flinch.
Free open source isn't your competitor. Assembly is.
The first reframe that helped: I'm not competing with the open-source MCP server. I'm competing with the customer spending their own time assembling and maintaining one. Those are different things.
The open-source tool is free to download. It is not free to run in production. Someone has to scope the role, stand up the proxy, add the logging, handle the credential storage, keep it patched, and own it when it breaks at an inconvenient hour. That someone is an engineer, and engineer-hours are the most expensive thing at most companies. The "free" option has a price tag. It's just denominated in time and attention instead of dollars, and it recurs forever.
So the real question a buyer faces isn't "free vs paid." It's "is the thing I'd build and maintain myself worth more or less than what they charge." For a solo hacker with time and curiosity, build it. For a team whose engineers have better things to do than maintain database-access plumbing, the math flips.
Sell the part that's annoying to own, not the part that's easy to build
The mistake I see founders make against open source is competing on the core feature, the part that's easy to replicate. The open-source server runs queries. Mine runs queries. If that's the battleground, I lose, because theirs is free and the query-running is genuinely not hard.
The things that are actually hard to own over time are different: the audit log that's there when an incident happens, the credential handling that doesn't leave secrets in a config file, the schema annotations that make the agent accurate, the permission model across multiple databases, the fact that when something breaks it's my problem and not the customer's weekend. None of those are the flashy core feature. All of them are the stuff that's tedious to build well and worse to maintain.
That's what you're actually selling against free: not the feature, the ownership. The customer isn't paying for query execution. They're paying to not be the one maintaining the thing that executes queries.
Don't be ashamed of the free path
Here's the part that felt counterintuitive: I think being honest about the free path makes the paid one easier to sell, not harder.
I publish the do-it-yourself instructions. I tell people the read-only role plus a timeout gets them most of the safety. Some readers do exactly that and never pay me, and that's fine, because they were never going to pay me. They have the time and the inclination. Hiding the free path wouldn't have converted them. It would just have made me look like I was pretending it didn't exist.
The people who do convert read the same free instructions, recognize that they could do it, and decide they'd rather not own it. That's a much better customer than someone who paid because they didn't realize there was a free option. The honesty filters for the people the product is actually for, and it builds the kind of trust that makes the eventual purchase feel obvious instead of extracted.
Where I landed on the actual price
I don't price against the free tool's price, which is zero, because that's a race I can't win and shouldn't try to. I price against the cost of a team doing it themselves and maintaining it: a fraction of an engineer's time, recurring. If the product saves more of that than it costs, the price is fair, and if it doesn't, no clever positioning will save it.
That's the discipline free open source actually imposes, and it's a healthy one. It forces you to be worth more than the time you save, which is the only honest reason to charge for software anyway. Free competitors don't kill you. They just make you prove your value isn't the thing that's easy to copy. For QueryBear, the value was never running the query. It was being the one who maintains everything around it so the customer doesn't have to.