---
name: vaneform-lookup
description: Look up cached website scale, registry, site profile, link-rank facts, Pro search footprint, keyword occupancy, and domain compares from Vaneform. Use when the user asks how big a site is, 查站, 体量, 关键词查询, 批量对比, 跨后缀对比, or remaining quota. Prefer the Vaneform MCP tools; otherwise use the matching GET /api/v1 path with the same vf_live_ key.
---

# Vaneform lookup

One `vf_live_` key. Same daily quota as the website. Do not put the key in this file.

## Install

Save this file as `.agents/skills/vaneform-lookup/SKILL.md` (project) or `~/.cursor/skills/vaneform-lookup/SKILL.md` (user). Connect MCP with the same key first.

## Prefer MCP

If the Vaneform MCP server is connected, call:

1. `get_account` before a batch — remaining lookups and the Pro search-footprint grant, no quota spend.
2. `lookup_domain` with `domain` and optional `include` (`scale,registry,profile,popularity` plus opt-in `search`).
3. `lookup_keyword` with `q` — Pro keyword brief (US/EN). Cache hit is free; a miss spends the shared advanced grant.
4. `compare_bulk` with `q` — comma-separated domains. Scale plus registry. One daily lookup per domain.
5. `compare_tld` with `q` and optional `s` — same name across suffixes. Default `com,ai,io`.

MCP URL: https://vaneform.com/mcp

If MCP is not connected, use the same key over HTTP:

```
GET https://vaneform.com/api/v1/domains/{domain}?include=scale,registry,profile,popularity
GET https://vaneform.com/api/v1/keyword?q=ai
GET https://vaneform.com/api/v1/bulk?q=google.com,apple.com
GET https://vaneform.com/api/v1/tld?q=google
GET https://vaneform.com/api/v1/account
Authorization: Bearer vf_live_…
```

Add `search` only for Pro search footprint: `include=scale,search` or `GET /api/v1/domains/{domain}/search`.

## How to read the result

- Domain `lookup_domain` is cache-only. `status=missing` means the cache has no snapshot. Do not invent visits, ranks, registry dates, or keywords.
- Every estimate keeps `as_of` and `confidence`. Quote them. Scale visits use `estimate`: `site` (full-site) or `search` (search-side).
- Domain `refresh_policy` is always `cache_only`. Do not ask that call to refresh or buy an upstream fill.
- `include` order is scale (体量), registry (注册局), profile (画像), popularity (外链), search (足迹). Popularity JSON uses `link_rank` / `board_rank`.
- `include=search` is Pro-only. Free callers get `search_requires_pro`. A successful read does not spend the 30-per-day footprint grant. `include_subdomains` is false: the queried host only.
- `lookup_keyword` is Pro-only (`pro_required` otherwise). Market is US/EN. Occupancy is Google depth 30.
- `compare_bulk` / `compare_tld` refresh scale and registry like the website tools. Free 3 / Pro 30.
- A successful `lookup_domain` or compare domain counts as one daily lookup. `get_account` does not. Keyword cache hits do not. Failures do not. Unpaid accounts also have a 5-per-5-minute pace.

## Output

Lead with the domain or keyword, status, and the numbers that exist. Link `report_url` when present. If the payload is an error, use its `error.code` (`unauthorized`, `invalid_api_key`, `lookup_limit_exceeded`, `lookup_rate_exceeded`, `search_requires_pro`, `pro_required`, `invalid_domain`, `invalid_request`).