Comparison
Slipdock vs GoDaddy's Domains API, for an AI agent
GoDaddy has thought harder about agent safety than anyone else selling domains, and it shows: a quote endpoint that returns a short-lived price token you must use to register, consent records capturing which agreements were accepted and when, a purchase-approval field the API will not let an agent skip, and scoped tokens you can revoke independently. Read that as validation, not as a competitor to sneer at. The gap is the same as everywhere else: a GoDaddy account with a payment profile has to exist before an agent can do anything, and there is still no spend cap — an approval flag is not a ceiling.
What GoDaddy gets right, and we copied the thinking
- A quote must precede a purchase. A short-lived price
token means the agent cannot register at a price nobody saw. Our equivalent
is that
GET /v1/domains/searchreturns a firm quote and renewal equals registration, so the number the agent showed its user stays true. - Consent is recorded, not assumed. Which agreements, and when. That is the right instinct for a flow where the buyer is software.
- Approval is a required field. The API refuses to proceed without it, so an agent cannot skip the step by omission — the strongest incumbent design we found.
- Scoped, independently revocable tokens. Also our model.
- The card never touches the API. Purchases draw from the payment profile on the account. Our wallet is the same idea with the ceiling added.
Where an agent starting from nothing still stops
| Step | GoDaddy | Slipdock |
|---|---|---|
| Credentials | API keys from a GoDaddy account. Creating the account is a browser flow. | One unauthenticated POST returns a scoped token. |
| Ability to buy | A payment profile must be configured; quotes fail if billing is not set up. | No payment method needed. The first purchase emails the owner. |
| Approval | A required field on the purchase call — strong, but set by the caller. | An out-of-band human approval: a link in the owner's inbox, on a page the agent cannot reach. |
| Spend ceiling | None in the API. | A per-purchase cap the human sets, enforced server-side. |
| Price | Retail. | At or near cost — .com $10.44. |
| Other resources | Hosting exists, bought separately. | Domain, DNS, git repo, app and Postgres under one token and one bill. |
Approval flag versus enforced cap
The distinction matters and is easy to blur. A required approval field means the agent must assert that approval happened. An enforced cap means the purchase cannot exceed an amount a human set, whatever the agent asserts. The first prevents an accident of omission; the second bounds a deliberate or manipulated request. GoDaddy has the first. We built the second, and we think both are worth having.
Use GoDaddy if…
- Your user is already a GoDaddy customer with billing configured.
- You need TLD breadth and aftermarket inventory today.
- You need something production-ready — they are, we are not yet.
Use Slipdock if…
- There is no account, no card and no browser available.
- You want a ceiling rather than a checkbox.
- You want the whole stack, not only the name.
Questions
Is GoDaddy's agent design better than Slipdock's?
In one respect, arguably: recorded consent artifacts per agreement are more thorough than ours today. In the two that decide whether an agent can start at all — no pre-existing funded account, and an enforced ceiling — we go further.
Are GoDaddy's prices higher?
They are retail rather than at cost, so generally yes for the popular TLDs. We would not make a decision on that alone; the difference on a single domain is small and their platform is mature.
Does GoDaddy have an MCP server?
Yes — their developer platform shipped one alongside the v3 Domains API, which is part of why their guardrail design is worth reading rather than dismissing.