Docs

For people. If you are an agent, the machine-readable version is llms.txt and openapi.json.

Quickstart

1 — Look something up. No account needed.

curl -s https://api.scopeweb.io/browse \
  -H 'content-type: application/json' \
  -d '{"query":"stowed","limit":5}'

Every row carries classification, rdap_source and last_verified_at. There is a daily allowance for anonymous callers and the response tells you what is left.

2 — Make an account when you want it to persist.

Sign in or create one — email and password, nothing to paste. Drafts, domains and watchlists live on the account. Your browser session authenticates the API directly.

3 — Price a name.

curl -s https://api.scopeweb.io/quote \
  -H 'content-type: application/json' \
  -d '{"domain":"example.com"}'

The breakdown is itemised: registration, WHOIS privacy where the registry sells it, and our service fee. Every line is a thing that was bought.

4 — Buying is a person's job.

A quote returns a confirm_url. That URL is not a purchase — it opens a page where a human types the domain to authorise the charge. There is deliberately no API call and no agent tool that completes an order.

Connect an agent

Point an MCP client at https://mcp.scopeweb.io/mcp. It registers itself, you approve the scopes on a consent screen, and that is the whole setup — nothing is issued to you in advance. See the connector page.

The endpoints

Rendered from the live spec, so this table cannot drift from the API.

tool	http	effect	what it does
create_draft	POST /drafts	writes	Store a draft website and get a persistent preview URL.
update_draft	PUT /drafts/{draft_id}	writes	Write a new immutable version of a draft (patch semantics).
list_drafts	GET /drafts	read-only	List the user's drafts.
get_draft	GET /drafts/{draft_id}	read-only	Read a draft's metadata, file list, or one file's contents.
verify_ownership	POST /domains/verify-{action}	writes	Prove domain ownership via a DNS TXT challenge.
list_domains	GET /domains	read-only	The user's verified domain inventory with cached scan data.
suggest_names	POST /browse	read-only	Suggest domain names for a concept. The first step when choosing a name.
score_name	POST /name/score	read-only	Grade candidate names on measurable, audited components.
scan_namespace	POST /scan	read-only	Classify an explicit list of domains.
check_live	POST /scan/live	read-only	Authoritative registration check. The purchase gate.
audit_domain	POST /scan/audit	read-only	Fetch one domain's live, observable surface.
set_budget	PUT /projects/{project_id}/budget	writes	Set a project's budget. Echoes carry and headroom immediately.
get_finance	GET /projects/{project_id}/finance	read-only	What a project costs to keep, and what is left.
kb_search	POST /kb/search	read-only	Search a curated corpus. Retrieval, not authority.
show_plans	GET /plans	read-only	The plans, what never varies between them, and where the caller stands.
quote_domain	POST /quote	read-only	Price an unregistered name. Proposing is allowed; spending is not.

The laws this API keeps

Unknown is never available. A lookup that failed is reported as unknown, never as a domain you can buy. Only an authoritative registry 404 says not-registered — and even that is not the same as purchasable.
Prices are identical for every buyer. They do not vary by budget, by plan, or by anything we know about you. See pricing.
Agents propose; people spend. No tool can complete a purchase.
We say what we did not measure. An unreadable value is reported as unknown rather than as zero, everywhere, including money.

Check us

The benchmark publishes accuracy against a fixed set with the harness command, so you can run it yourself. Coverage lists every TLD we can answer authoritatively and every endpoint that failed admission.

FOR AGENTS

llms.txt carries the interpretation rules; openapi.json is the machine surface; spec.json is the single source both are generated from.