mcpField
N
namecatcher-skills

namecatcher-skills

Local
github.com/productmapFinance0Updated 15w ago

About namecatcher-skills

NameCatcher provides username scoring and valuation for Telegram collectible names, for investors and AI agents comparing candidates. It combines a namability score, estimated fair price, and positioning thesis through an API, Agent Skill, and MCP server.

Key features

  1. 01Namability score from 0–100
  2. 02Fair price estimate in GRAM
  3. 03Positioning thesis for names
  4. 04Compare and rank candidates
  5. 05Check Fragment asking prices

Use cases

  1. 01Compare Telegram username candidates
  2. 02Assess a name before purchase
  3. 03Check an asking price against valuation

MCP support

The MCP server exposes tools to score one name, rank up to 20 candidates, and compare a Fragment asking price with the estimated valuation. It is available as a stdio bridge or hosted Streamable HTTP endpoint.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"productmap-namecatcher-skills": {
"command": "npx",
"args": [
"-y",
"namecatcher-mcp"
],
"env": {
"NAMECATCHER_API_ORIGIN": "<NAMECATCHER_API_ORIGIN>"
}
}
}
}
Run commandnpx -y namecatcher-mcp
Environment variables to set
NAMECATCHER_API_ORIGINOverride the NameCatcher API origin (optional; defaults to the public API)

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmnamecatcher-mcp0.1.1

Tools

3

Saved tool list · updated 2w ago

ParameterTypeRequiredDescription
langenumNoThesis language, default enen | ru
namestringYesUsername without @ (4-32 chars, a-z 0-9 _)

Details

From the project README

NameCatcher — username intelligence for AI agents

Score and value collectible Telegram usernames: namability score (0–100) validated against 111k+ real Fragment sales, fair price estimation in GRAM, and positioning theses. Free public API — no key, no wallet, no setup.

Three ways to plug in, thinnest first:

LayerForInstall
Plain HTTPany agent that can curl/fetchnothing
Agent SkillClaude Code / skills-compatible agentsnpx skills add productmap/namecatcher-skills
MCP serverClaude Desktop/Code, any MCP clientnpx -y namecatcher-mcp

Built by NameCatcher — the Telegram mini app for username investors: mint catching, market analytics, portfolio, value passports.

Quick try

curl -s "https://api-namecatcher.rocketname.com/public/score/wearlab?lang=en" | jq
{
  "username": "wearlab",
  "score": 67,
  "band": "strong",
  "fair_price_ton": 400,
  "thesis": "A crisp fusion of \"wear\" and \"lab\" …",
  "lang": "en",
  "analysis_url": "https://t.me/NameCatcherBot?startapp=name_wearlab__pub_api",
  "fragment_url": "https://fragment.com/username/wearlab"
}

API reference

One endpoint, no auth:

GET https://api-namecatcher.rocketname.com/public/score/{name}?lang={en|ru}
Parameters
ParamInRequiredDescription
namepathyesUsername without @. 4–32 chars, a-z 0-9 _
langquerynoen (default) or ru — language of thesis
Response fields
FieldTypeDescription
usernamestringCanonical (lowercase) username
scoreint 0–100Namability — quality of the name as a brandable asset. Market-validated: names scoring 90+ sell for a multiple of 80–89 names
bandstringpremium (≥80) · strong (≥66) · medium (≥50) · weak (<50). Below 50 is not investment-grade
fair_price_tonnumber | nullClass-based fair price estimate in GRAM (length × score class)
thesisstring | nullShort positioning summary (why the name works, for whom). Present only for names already analyzed — null is not a quality signal
langstringLanguage the thesis was returned in
analysis_urlstringFull breakdown in the NameCatcher mini app
fragment_urlstringThe name's page on Fragment (live market status, buy/bid)
Errors
StatusBodyMeaning
400{"error":"INVALID_NAME"}Not a valid username string
400{"error":"NOT_COLLECTIBLE"}Valid string, but not scorable as a collectible (e.g. too short)
429{"error":"RATE_LIMITED","retry_after_s":n}Over 60 req/min per IP — wait retry_after_s and retry
429{"error":"DAILY_LIMIT","limit_per_day":1000}Over 1,000 req/day per IP
Rate limits

60 requests/min and 1,000 requests/day per IP. Generous for agent workflows (evaluating and comparing dozens of names); not enough for bulk scraping — that's intentional. Need more for a legitimate use case? Ping @BrandEvangelist.

curl cookbook
# Score one name (Russian thesis)
curl -s "https://api-namecatcher.rocketname.com/public/score/morya?lang=ru" | jq '{score, band, fair_price_ton}'

# Compare candidates: pick the best of three
for n in pulsefit traintrack fitwave; do
  curl -s "https://api-namecatcher.rocketname.com/public/score/$n" | jq -c '{username, score, band}'
done | sort -t: -k2 -rn

# Pre-purchase check: is a 300 GRAM ask below class valuation?
curl -s "https://api-namecatcher.rocketname.com/public/score/vault" \
  | jq '{score, fair_price_ton, below_class: (.fair_price_ton / 300 >= 1.2)}'

Agent Skill

The names/ton-usernames skill teaches an agent the full workflow: single-name evaluation, candidate comparison, and pre-purchase checks against Fragment asks.

npx skills add productmap/namecatcher-skills --skill ton-usernames
# or manually:
cp -r names/ton-usernames ~/.claude/skills/

A PR adding this skill to the official ton-org/skills is open.

MCP server

A thin stdio bridge over the same API — for Claude Desktop, Claude Code, and any MCP client. No key needed. Published on npm as namecatcher-mcp (GitHub install npx -y github:productmap/namecatcher-skills also works).

smithery badge

Remote endpoint (Streamable HTTP) — same three tools, zero install:

https://mcp-namecatcher.rocketname.com/mcp
claude mcp add --transport http namecatcher https://mcp-namecatcher.rocketname.com/mcp

Claude Code:

claude mcp add namecatcher -- npx -y namecatcher-mcp

Claude Desktop / other MCP clients (mcpServers config):

{
  "mcpServers": {
    "namecatcher": {
      "command": "npx",
      "args": ["-y", "namecatcher-mcp"]
    }
  }
}
Tools
ToolDescriptionArguments
score_nameScore one username: namability, band, fair price, thesisname, lang?
compare_namesScore up to 20 candidates and rank them by scorenames[], lang?
check_fragment_askCompare a Fragment asking price to the class valuation → fair/ask ratio + verdictname, ask_ton

Environment: NAMECATCHER_API_ORIGIN (optional) — override the API origin.

Notes for builders

  • The score measures naming quality, deterministic per model version — safe to cache
  • It is not demand for a specific lot: always check the live Fragment price before money decisions
  • Pairs naturally with TON wallet tooling (@ton/mcp): evaluate → then buy

Links

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.