Comparison
Slipdock vs Cloudflare Registrar, for an AI agent
On price, Cloudflare wins and will keep winning: they resell .com at their own wholesale $10.44, which is below what our registrar charges us, so we charge $12.49. Their API also ships inside the Cloudflare MCP, so a lot of agents already have it installed and it takes zero integration work. The difference is the starting condition: Cloudflare's registrar API requires a Cloudflare account with a default payment method already on file, and it has no spend cap. Slipdock starts from nothing and enforces a cap. If your user already has that Cloudflare account and only wants a domain, use Cloudflare.
What Cloudflare is genuinely better at, today
- It exists in production. Slipdock does not yet. That is the single biggest difference on this page and no amount of design argument changes it.
- Distribution. The Registrar API is exposed through the Cloudflare MCP, already installed in a lot of agent toolchains. There is nothing to add and nothing to learn.
- Everything else they run. If the domain is going to sit behind Cloudflare's edge, WAF, Workers and CDN anyway, keeping the registrar in the same account is obviously simpler.
- At-cost pricing, permanently. They can do it because domains are a rounding error next to their real business, and because their wholesale price is below ours. We cannot match it without selling at a loss, so on price alone they win.
What breaks for an agent starting from nothing
| Step | Cloudflare Registrar | Slipdock |
|---|---|---|
| Get credentials | A Cloudflare account must already exist. Creating one is a browser flow. | One unauthenticated POST /v1/workspaces returns a scoped token. |
| Be able to buy | A default payment method must already be on the account. | Nothing needed up front. The purchase emails the owner a link. |
| Limit what the agent can spend | No spend cap in the API. Their documentation states that designing a flow which will not buy domains without approval is the developer's responsibility. | A per-purchase cap the human sets, enforced server-side, with no card on file to fall back on. |
| Nameservers | Domains registered there use Cloudflare's DNS. | Our authoritative DNS, free, or point the zone wherever you like. |
| Other resources under the same token | Cloudflare's own products. | Domain, DNS, private git repo, app hosting and Postgres. |
The spend-cap point, stated fairly
Cloudflare is not careless here; they are explicit. Their registrar documentation says in plain language that it is the human's responsibility to design an agent flow that will not buy domains without approval. That is an honest disclosure and a sensible position for a platform whose customers are developers.
It is also a guarantee that lives in your prompt rather than in their server, and prompts are not a security boundary. Slipdock's position is that the cap belongs in the infrastructure, because that is the only place it can be enforced rather than requested. Reasonable people can disagree; the difference is worth knowing before you pick.
Price, with nothing hidden
| Cloudflare | Slipdock | |
|---|---|---|
| .com | $10.44 | $12.49 |
| Renewal | same as registration | same as registration |
| Our wholesale cost | lower than ours, and they add nothing | $11.50 + $0.66 card fee |
| What we keep | n/a | $0.33/yr |
| Where the margin is | not the point for them | hosting, databases and the wallet |
Cloudflare is cheaper and will stay cheaper. They buy .com for less than we do and resell it at exactly that, so matching them would mean losing money on every domain — and a hook paid for out of our own runway is one an abuser can farm. If a domain is all you need, use Cloudflare; we would rather say that than pretend a $2.05 difference is a feature.
What the difference buys is the rest of the token: DNS, TLS, a git repo, hosting and a database behind one credential your agent gets in its first second, with a spend cap the API enforces. If you want only the domain, that is not worth $2.05 to you, and it should not be.
The nameserver condition, and leaving generally
A domain registered with Cloudflare Registrar uses Cloudflare's DNS. That is a reasonable consequence of at-cost pricing — the domain is a hook for the platform, and the hook is the point — but it is a condition attached to the price, and it is worth naming. It is also, incidentally, the trade we chose not to make: rather than tie you to our nameservers to justify a price below our cost, we charge a little over our cost and leave the zone where you want it. Slipdock's authoritative DNS is a free default, not a requirement: point the zone anywhere you like and the price does not change.
More broadly, we publish a written commitment that leaving is easy: repo, database, DNS zone and billing history all export in standard formats, at any time, for no fee, including while a workspace is suspended for non-payment. In fairness, domains transfer out of Cloudflare too — ICANN requires that of everyone — so the difference is scope rather than the domain itself. The commitment.
Use Cloudflare if…
- Your user already has an account with billing configured.
- The domain is going behind Cloudflare's edge anyway.
- You need something that works in production this week.
Use Slipdock if…
- There is no account and no card, and there is no human at the keyboard right now.
- You want the ceiling on agent spending to be enforced by the server rather than by your prompt.
- You want the domain, the DNS, the repo and the running app to be one token and one bill.
Questions
Is Slipdock cheaper than Cloudflare?
No, and we cannot be. Cloudflare resells .com at their own wholesale price of $10.44, which is below what our registrar charges us ($11.50). We charge $12.49 — registrar, card fee, and $0.33 for us. Price is not the reason to choose us.
Can Cloudflare's API be used by an agent at all?
Yes, and well, once the account and payment method exist. The limitation is the cold start and the absence of an enforced cap, not the API surface.
What if Cloudflare ships registrar-as-a-service?
Then the domain leg of this comparison largely closes, which we think is likely and have said so internally. Our answer is that the domain was never the product: the token, the enforced mandate and the aggregation across domains, DNS, repos, apps and databases are.