namescope
MCP server for evaluating software project name availability across package registries, GitHub, domains, and the command line, with scoring and evidence.
README
NameScope
Picking a project name is easy. Finding out whether that name is usable across package registries, GitHub, domains, and the command line is the tedious part.
NameScope does that research from your terminal. Give it a project description or a shortlist of names and it returns a scored report with the evidence behind every result. Generation and scoring run locally; network checks use public provider endpoints.
NameScope is built for software projects. It cannot promise that a business name, domain, social handle, or trademark is legally available.
Quick start
NameScope requires Node.js 24 LTS or newer. You can run it without installing anything globally:
npx namescope find "a fast open-source database migration tool"
npx namescope check taskforge
npx namescope rank taskforge taskmint orbitdesk
No account, subscription, payment, hosted backend, Docker setup, or language model is required. A GITHUB_TOKEN is optional and only increases GitHub API rate limits.
How it works
find starts with a project description and generates deterministic candidates. check starts with one name, while rank compares several. All three eventually use the same checking and scoring engine.
flowchart TD
A["Project description or candidate names"] --> B{"Command"}
B -->|find| C["Generate candidates locally"]
B -->|check or rank| D["Use supplied names"]
C --> E["Normalize query variants"]
D --> E
E --> F["Run local quality and CLI checks"]
E --> G{"Offline mode?"}
G -->|Yes| H["Skip remote requests and record unknowns"]
G -->|No| I["Query public providers in parallel"]
I --> J["RDAP, npm, PyPI, crates.io, and GitHub"]
F --> K["Collect evidence, confidence, and timestamps"]
H --> K
J --> K
K --> L["Add optional trademark links and disclaimer"]
L --> M["Calculate weighted dimension scores"]
M --> N{"Blocking conflict?"}
N -->|Yes| O["Cap or block the recommendation"]
N -->|No| P["Keep the composite score"]
O --> Q["Return human, JSON, Markdown, or MCP output"]
P --> Q
Provider-specific commands such as domains and packages return their evidence directly instead of building a complete composite report.
1. Generate or accept candidates
Name generation is local and reproducible. NameScope combines useful words from the description with related terms, compounds, prefixes, suffixes, and blends. Generated names are only ideas; they do not carry availability claims until checks run.
2. Check public and local namespaces
NameScope checks each candidate against the sources relevant to software projects:
| Area | What is checked | Important limitation |
|---|---|---|
| Packages | npm name validity, then exact and separator variants on npm, PyPI, and crates.io | Registry state can change immediately |
| GitHub | Repository-name search and exact user or organization namespace | Search is not proof of global availability |
| Domains | Selected TLDs through public RDAP | A 404 is a useful signal, not a purchase guarantee |
| CLI | Local PATH, common commands, language tools, and shell-reserved names |
Results describe the current machine |
| Name quality | Length, word count, pronunciation proxy, punctuation, numbers, ambiguity, and relevance | These are heuristics, not user research |
| Trademark | Official manual-search links and a legal disclaimer when requested | This is not trademark clearance or legal advice |
Search distinctiveness stays neutral until a search adapter exists. Social handles are not checked in version 0.1.
3. Keep uncertainty visible
A timeout, rate limit, blocked request, unsupported provider, or malformed response remains unknown, unsupported, or error. NameScope never turns a failed lookup into “available.”
Every completed report includes provider evidence, confidence, timestamps, warnings, unknown checks, dimension scores, weights, and a plain-language explanation.
4. Score and rank
The default score combines package uniqueness, GitHub uniqueness, domain options, deterministic name quality, search distinctiveness, and CLI usability. Weights are configurable and normalized by their actual sum, so they do not need to total 100.
A package collision or another confirmed blocking conflict can cap or block a recommendation. A high score still means “strong candidate based on these checks,” not “safe everywhere.”
Commands
| Command | Use it to |
|---|---|
namescope find <description> |
Generate, check, and rank project-name ideas |
namescope check <name> |
Build a complete report for one name |
namescope rank <name...> |
Compare two or more existing candidates |
namescope domains <name> |
Check selected domain extensions only |
namescope packages <name> |
Check selected package registries only |
namescope mcp |
Start the local MCP stdio server |
Useful examples:
# Check selected domains
npx namescope domains taskforge --tlds com,io,dev,app
# Check selected package registries
npx namescope packages taskforge --registries npm,pypi,crates
# Return structured output
npx namescope check taskforge --json
# Prevent every network request
npx namescope check taskforge --offline
# Include preliminary trademark resources
npx namescope check taskforge --include-trademark --nice-classes 9,42
Output formats are human-readable text, --json, --compact-json, and --markdown.
Privacy and offline use
Generation, normalization, quality analysis, scoring, cache handling, and CLI collision checks run locally. Remote commands send only the requested candidate and query variant to the selected providers.
Use --offline to prevent all network requests. NameScope has no telemetry and does not collect project names or descriptions. Read the privacy documentation for provider-by-provider details.
Configuration
Copy namecheck.config.example.json to namecheck.config.json. Use it to change TLDs, registries, scoring weights, timeouts, cache lifetime, concurrency, or excluded words. CLI flags override file values.
See the configuration guide and scoring methodology.
MCP setup
NameScope exposes the same engine through a local MCP stdio server:
{
"mcpServers": {
"namescope": {
"command": "npx",
"args": ["-y", "namescope", "mcp"]
}
}
}
Available tools are generate_names, check_name, rank_names, check_domains, check_packages, check_github, check_trademark, and explain_score. Read the MCP guide for details.
Development
git clone https://github.com/ensp1re/namescope.git
cd namescope
npm install
npm run check
npm run build
node packages/cli/dist/index.js find "an open-source TypeScript job queue" --offline
The repository is an npm workspace split into schemas, core logic, provider adapters, MCP, and CLI packages. Read the architecture guide for the dependency flow and CONTRIBUTING.md before sending a change.
Know the limits
Results are snapshots, and public namespaces can change at any time. RDAP behavior varies between registries, GitHub search is not exhaustive, local command checks vary by machine, and deterministic quality scores cannot replace research with real people.
Trademark output is preliminary information, not legal advice or a comprehensive clearance search. Consult a qualified trademark attorney before relying on a name commercially. The full list lives in limitations.
License
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.