# NameScout

what it does

type a project name. get back one report covering everywhere a collision could trip you up:

• domains: .com, .dev, .io, .app, .ai, .co
• package registries: npm, pypi, rubygems, packagist, homebrew, docker hub
• code hosts: github user/org, github repo search
• plugin marketplaces: vs code, obsidian (coming), chrome web store (coming)

each provider returns taken / available / unknown. a mode-weighted score (0–100) rolls everything into one verdict: mostly clear, usable but watch it, or collision risk. modes shift the weights: saas prioritizes domains, library prioritizes packages, developer tool prioritizes both.

if the name is taken, the report also surfaces near-matches across npm, github, and pypi so you can see what people might confuse you with, plus an opt-in panel of ai-suggested alternatives.

why i built it

every time i start a new project i run through the same routine: check the .com, check the .dev, check npm, check github, check whether someone already shipped a tool with this name three years ago. it takes 10 minutes and i forget at least one source every time. by the time i’ve named the repo, half the namespaces are already gone.

namescout is that routine collapsed into one query. the goal isn’t to be clever. it’s to never have to remember which 12 places to look.

the cli

namescout also ships as a one-command cli so you can run a check without leaving the terminal:

npx namescout-cli redis

same checks, same scoring, terminal-pretty output. --mode, --suggest, --json, and --no-color flags supported. the cli has its own tool page → with install + flag docs.

how it’s built

a quick tour:

• next 16 app on vercel, react 19, tailwind 4
• per-provider registry: every check is one { id, check(name) } entry. adding a provider is a single array append, not a refactor
• streaming via suspense: page shell renders synchronously at ~40ms ttfb; each provider fans out in its own suspense boundary. react cache() dedupes the shared fetcher across the report so 11 components don’t each refetch
• dynamic og images at /og: per-query card with score, verdict, and the top taken providers, generated server-side via next/og
• /api/v1/* endpoints: check, suggest, near-matches, all open with 30 req/min per ip and standard x-ratelimit-* headers
• ai-assisted suggestions: opt-in ?suggest=1 panel, openai backed with a programmatic fallback. cached per (query, mode, count) for an hour so the same name doesn’t burn tokens twice

what’s next

• premium watch feature: pick a name, get pinged when a key namespace changes (a .dev opens up, an npm package is unpublished, etc.). $5/mo, the one piece of real recurring revenue this tool can sustain
• registrar affiliate integration: when a domain is available, link through to a registrar with an affiliate id. passive monetization that doesn’t get in the user’s way
• open-sourcing the main repo once the watch feature ships. the cli is already public on github

the next thing is making the report run a watch in the background, so “this name is gone forever” turns into “i’ll know the moment it isn’t.”