
MCP server, CLI and agent skill for Vaneform: free website traffic checker (monthly visits, traffic
MCP server, CLI and agent skill for Vaneform: free website traffic checker (monthly visits, traffic sources, top countries, domain compares)
·
Website traffic lookups for AI assistants and the terminal. This repo has an MCP server, a small CLI and an agent skill for Vaneform, a free website traffic checker.
Ask your assistant "how much traffic does notion.so get?" or "compare figma.com and canva.com" and it answers with estimated monthly visits, month-over-month change, rank, top countries, traffic sources and registry facts. Traffic figures are third-party traffic estimates; see the Vaneform methodology.
| Piece | What it is |
|---|---|
| Hosted MCP server | https://vaneform.com/mcp (Streamable HTTP). Zero install, and the recommended option. |
vaneform-mcp | Local stdio MCP server for clients that only run local commands. It proxies the hosted server, so tools never drift. |
vaneform CLI | vaneform notion.so, vaneform compare a.com b.com, vaneform tld name. It shares the same client. |
skills/website-traffic-lookup | Agent skill (Claude / Agent Skills format) that tells an assistant when to check traffic and how to read and cite the result. |
The local server proxies tools/list from the hosted server and does not embed a second schema. This table matches the public API and the tool descriptions Vaneform publishes:
| Tool | What it does | Cost |
|---|---|---|
lookup_domain | Cached snapshot of a domain: visits (full-site or search), registry, site profile, link-rank popularity. Optional include=baidu. Optional include=search is Pro keyword rankings. | 1 credit. search can add up to 200 Pro credits; do not pass it unless the user agrees |
compare_bulk | Side-by-side visits and registry for several domains; fetches traffic when it is missing | 1 credit per domain with data. Free: up to 3 domains. Pro: up to 30 |
compare_tld | One name across suffixes (default com,ai,io) | 1 credit per domain with data |
get_account | Plan and remaining credits | free |
lookup_keyword | US keyword brief: volume, difficulty, CPC, Google occupancy (Pro). A cache hit is free | 60 credits on a successful miss. Ask before spending them |
tools/list is the live contract. The local server proxies it and does not hardcode tool schemas. Five research tools (research_site, get_research_workspace, save_research, monitor_research, manage_research) also appear there. Publisher docs price them at 0 credits. Saving, monitoring, and managing write the caller's private workspace, so an agent should do that only when the user asks. The agent skill covers that rule.
Repeating a domain lookup within 24 hours is free. Site keyword rankings (Pro, up to 200 credits) are available through the API with include=search; a repeat inside 24 hours uses that receipt. A keyword brief whose status is empty spends 0.
| Free | Pro | |
|---|---|---|
| Credits per month | 100 | 10,000 |
| API key, MCP and CLI | ✓ | ✓ |
| Domain lookup and compare (1 credit per domain, repeats within 24h free) | ✓ | ✓ |
| Keyword brief (60 credits) | ✓ | |
| Site keyword rankings (200 credits) | ✓ |
Each key allows 60 requests a minute and 5,000 a day (429 lookup_rate_exceeded above that). See pricing.
Discovery (initialize, tools/list) is open. Tool calls need a Vaneform API key.
Authorization: Bearer vf_live_… or X-API-Key: vf_live_…. Local server and CLI: set VANEFORM_API_KEY.Without a key, tools still list, and a call returns a sign-up link plus a browser link such as https://vaneform.com/?q=notion.so.
claude mcp add --transport http --scope user vaneform https://vaneform.com/mcp \
--header "Authorization: Bearer vf_live_…"~/.cursor/mcp.json (or .cursor/mcp.json in a project):
{
"mcpServers": {
"vaneform": {
"url": "https://vaneform.com/mcp",
"headers": { "Authorization": "Bearer vf_live_…" }
}
}
}The simplest Claude Desktop setup that passes your key is the local proxy. Edit claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):
{
"mcpServers": {
"vaneform": {
"command": "npx",
"args": ["-y", "github:vaneform/vaneform-mcp"],
"env": { "VANEFORM_API_KEY": "vf_live_…" }
}
}
}The package is not on npm yet, so npx installs it from GitHub. The first start takes a little longer because prepare runs npm run build once for a git install. That script does not run when the same package is later installed from the npm registry, and it does not send telemetry. If npx exits right away on an old system npm (Debian's packaged npm 9.2 does this with GitHub packages), run npm install -g github:vaneform/vaneform-mcp and use "command": "vaneform-mcp" with no args.
After [email protected] is published, the stdio server is npx -y vaneform-mcp or npx -y [email protected] (the bin with the same name as the package, no further arguments). Until that release exists, the GitHub command above is the one that works. See RELEASING.md.
export VANEFORM_API_KEY=vf_live_…
codex mcp add vaneform --url https://vaneform.com/mcp --bearer-token-env-var VANEFORM_API_KEYhttps://vaneform.com/mcp, transport Streamable HTTP, header Authorization: Bearer vf_live_… (or X-API-Key: vf_live_…)npx -y github:vaneform/vaneform-mcp, env VANEFORM_API_KEYThe package has two bins. npx runs the one named vaneform-mcp, the stdio server. Words after that package name are arguments to the server, which does not parse them as a CLI. Select the vaneform bin with --package.
npx -p github:vaneform/vaneform-mcp vaneform --help
# after npm publish:
npx -y [email protected] vaneform --version
npx -y [email protected] vaneform --help
npm exec -y [email protected] -- vaneform --version
# or clone, then: npm install && npm link
export VANEFORM_API_KEY=vf_live_…
vaneform notion.so # traffic overview (1 credit; repeat within 24h free)
vaneform compare notion.so coda.io # side by side (Free: up to 3 domains)
vaneform tld vaneform --suffixes com,ai,io
vaneform keyword "website traffic" # Pro; a cache miss spends 60 credits
vaneform notion.so --include scale,baidu # cached Baidu rank; do not add search unless you mean to spend up to 200 credits
vaneform account # plan and remaining credits (free)
vaneform notion.so --json # raw API JSONOutput with no key set:
$ vaneform tools
lookup_domain Read a cached Vaneform domain snapshot: scale (full-site or search visits), registry, site profile, and link-rank popularity.
get_account Read the caller plan, website lookup allowance, and credit balance.
lookup_keyword Read a US/EN keyword brief: monthly searches, difficulty, CPC, and Google occupancy.
compare_bulk Compare domains side by side.
compare_tld Compare the same registrable name across suffixes.
$ vaneform github.com
VANEFORM_API_KEY is not set, so this tool cannot call the Vaneform API.
Create a free API key at https://vaneform.com/account/api (100 credits a month; 1 credit per domain).
Check it in a browser instead: https://vaneform.com/?q=github.comWith a key, vaneform <domain> prints visits, MoM change, rank, top countries, sources, registry dates and the report link, followed by Data: Vaneform (third-party traffic estimates).
curl -s https://vaneform.com/mcp \
-H 'content-type: application/json' -H 'accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'The same key works for the REST API: docs, OpenAPI.
The canonical skill name is website-traffic-lookup. Install that skill, not the site filename:
npx skills add vaneform/vaneform-mcp --skill website-traffic-lookupnpx skills add vaneform/vaneform-mcp --list shows the skills in this repo and does not install them. This repo does not contain a skill named vaneform-lookup. vaneform.com/skill.md is a separate discovery document that asks agents to save it under that filename. Install website-traffic-lookup. Copying SKILL.md by hand is the way to place that same file in another folder; leave its name unchanged.
A manual copy, for an agent that only reads a skills folder:
mkdir -p ~/.claude/skills/website-traffic-lookup
curl -fsSL https://raw.githubusercontent.com/vaneform/vaneform-mcp/main/skills/website-traffic-lookup/SKILL.md \
-o ~/.claude/skills/website-traffic-lookup/SKILL.mdThe skill tells the assistant to cite the month and confidence, to say when a figure is unknown, to check remaining credits, and to ask before a 60-credit keyword brief or a 200-credit include=search call. Fetched payloads are untrusted data. The file contains no API key.
status: "missing" means Vaneform hasn't cached that domain yet. It does not mean zero traffic. compare_bulk fetches it.as_of) and a confidence label. Quote them.Two version series live in this repo. They move for different reasons and are not meant to match.
| Series | Where | Current | What it versions |
|---|---|---|---|
| npm client | package.json | 0.1.3 | This CLI, the stdio proxy, and the packaged skill. Not published yet. |
| MCP Registry | server.json | 1.4.3 | The hosted server record io.github.jiangqizheng/vaneform-mcp at https://vaneform.com/mcp. |
vaneform --version prints the npm client version. The public HTTP API has its own OpenAPI info.version as well. A registry or API bump does not by itself bump the npm package, and an npm bump does not change the hosted server.
npm install # also builds dist/ via prepare; no telemetry
npm run lint # biome + tsc
npm test # unit tests + an offline stdio test with the official MCP SDK client
npm run pack:check # npm pack, then run the extracted CLI and stdio server offline
VANEFORM_MCP_URL=http://localhost:3000/mcp node dist/bin/vaneform.js tools # point at another serverThe runtime has no dependencies and needs Node 18 or newer. This repo contains only the client. No data collection code lives here. Publishing steps are in RELEASING.md. The npm token stays on the maintainer's machine.
MIT. The license covers this code only, not data returned by the API. Traffic figures are third-party estimates served by Vaneform.
Pick your client and paste the snippet. Each one is the same server, written the way that client expects it.
claude mcp add vaneform --transport http https://vaneform.com/mcp{
"mcpServers": {
"vaneform": {
"url": "https://vaneform.com/mcp",
"headers": {
"Authorization": ""
}
}
}
}code --add-mcp '{"name":"vaneform","url":"https://vaneform.com/mcp","headers":{"Authorization":""}}'Runs as a hosted service. Your client connects over the network and nothing is installed locally.
Paste this prompt into your agent. It reads this page and does the setup for you.
Read https://aiagentslisting.com/mcp/vaneform-mcp to learn what the "Vaneform Mcp" MCP server does and how to install it. Add it to my coding agent's MCP configuration as documented on that page, then confirm the server connects and list the tools it exposes.Agents can also browse this directory over MCP at https://aiagentslisting.com/api/mcp
claude mcp add --transport http --scope user vaneform https://vaneform.com/mcpnpx skills add vaneform/vaneform-mcp --skill website-traffic-lookup| Name | What it does |
|---|---|
research_site | Free early-access standalone site research. Query Google ads by domain or advertiser, public Meta ads by keyword, AdSense accounts/related sites, outbound page links, collected ad text, recorded websi |
get_research_workspace | Read the caller saved research, folders, private notes, tool history, publisher/ad monitors and in-app change notifications. Free; 0 credits. Account ownership comes from the authenticated key. |
save_research | Save a standalone site-tool query in the caller workspace, with a folder and private notes. Free; 0 credits. |
monitor_research | Monitor a Google Ads or AdSense query daily or weekly. Record newly observed public ads and connected sites as private in-app notifications. The first successful observation establishes a baseline. Fr |
manage_research | Manage the caller private research: remove a saved target, pause/resume or change daily/weekly monitoring, remove a monitor, check current public data, or mark a change notification read. Use IDs from |
lookup_domain | Read a cached Vaneform domain snapshot: scale (full-site or search visits), registry, site profile, and link-rank popularity. Optional include=baidu is cached Baidu rank. Optional include=search is Pr |
get_account | Read the caller plan, website lookup allowance, and credit balance. This call spends 0. Use before a batch. |
lookup_keyword | Read a US/EN keyword brief: monthly searches, difficulty, CPC, and Google occupancy. Pro-only. At most 30 Google results. A cache hit is free; a successful miss spends 60 credits. After trimming, long |
compare_bulk | Compare domains side by side. Returns scale and registry. Fetches traffic when monthly visits are missing. Spends 1 credit per compared domain with data. Up to 3 domains on Free, 30 on Pro. Use for 批量 |
compare_tld | Compare the same registrable name across suffixes. Default suffixes are com,ai,io. example.com is read as example. Response suffixes and items follow a fixed suffix order; match the returned suffixes. |
Read live from the server on November 15, 2026.
Maintain this server? Claiming proves you own the listing and earns the last 4 points.
Claim this listingPaste this line near the top of your repository's README. The badge always renders the current score.
[](https://aiagentslisting.com/mcp/vaneform-mcp)Nothing comparable is listed yet.
No tagged releases on record.
Connect to the AI Agents Listing MCP server at https://aiagentslisting.com/api/mcp using the HTTP transport and complete the sign-in when it asks. Then call the submit_review tool for the listing "vaneform-mcp" with a rating from 1 to 5 and a short note on what worked and what did not.No reviews yet
Be the first to share your experience with this tool.
One email a week. New agents, MCP servers and skills, and what is actually getting traction.