Internet-Names-MCP
Check availability of domain names, social media handles and subreddits
Documentation
Internet Names MCP Server
An MCP server for checking availability of domain names, social media handles, and subreddits. Returns clean JSON responses suitable for programmatic use.
Features
- Domain names - Check availability via RDAP (free) or NameSilo API (free, but requires API key - responses include domain prices too)
- Social media handles - Instagram, Twitter/X, Reddit, YouTube, TikTok, Twitch, Threads, Bluesky, GitHub, Snapchat, Pinterest, Kick, Substack
- Subreddits - Check if subreddit names are available on Reddit
- Comprehensive search - Generate name combinations and check everything at once
Quick Start
1. Add to Claude Code (Recommended)
claude mcp add --scope user internet-names-mcp uvx internet-names-mcp
That's it! The server works immediately using RDAP for domain lookups.
2. Optional: Configure NameSilo API (for domain pricing)
Checking domain name availability works best if you configure a NameSilo API key. It's free and requires just a basic registration. Otherwise we fall back to using RDAP, which has some significant limitations (see below).
Set up your NameSilo API key
- Create an account at namesilo.com (or log in)
- Go to API Manager
- Click Generate New API Key
- Copy the key and run:
uvx internet-names-mcp --setup
You'll be prompted to paste your API key.
On macOS, your API key is stored safely in your iCloud or login keychain as internet-names-mcp.namesilo. On other platforms, it's stored in ~/.config/internet-names-mcp/config.json.
Using --setup is the preferred method as it stores the key securely. Alternatively, you can set it via environment variable (see Manual Configuration below).
Manual Configuration
If you're not using Claude Code, or prefer to configure MCP servers manually, add this to your MCP client's configuration file (e.g., claude_desktop_config.json for Claude Desktop):
{
"mcpServers": {
"internet-names-mcp": {
"command": "uvx",
"args": ["internet-names-mcp"]
}
}
}
To include a NameSilo API key via environment variable (if you haven't used --setup):
{
"mcpServers": {
"internet-names-mcp": {
"command": "uvx",
"args": ["internet-names-mcp"],
"env": {
"NAMESILO_API_KEY": "your_api_key_here"
}
}
}
}
Note: Using uvx internet-names-mcp --setup to store the API key is preferred over environment variables, as it uses secure storage (macOS Keychain) when available.
CLI Commands
uvx internet-names-mcp --setup # Configure API keys interactively
uvx internet-names-mcp --show-config # Show current configuration
uvx internet-names-mcp --version # Show version
uvx internet-names-mcp --help # Show help
Tools
get_supported_socials()
Returns list of supported social media platforms.
Response:
{
"platforms": ["instagram", "twitter", "reddit", "youtube", "tiktok", "twitch", "threads", "bluesky", "github", "snapchat", "pinterest", "kick", "substack", "subreddit"]
}
Note: subreddit is checked via check_subreddits(), not check_handles().
check_domains(names, tlds?, method?, only_report_available?)
Check domain name availability and pricing.
Parameters:
| Name | Type | Default | Description |
|---|---|---|---|
names | list[str] | required | Domain names or base names to check |
tlds | list[str] | ["com", "io", "ai", "co", "app", "dev", "net", "org"] | TLDs to check |
method | str | "auto" | "auto", "rdap", or "namesilo" |
only_report_available | bool | false | If true, omit unavailable domains from response |
If a name contains a dot, it's treated as a full domain. Otherwise, it's combined with each TLD.
Response:
{
"available": [
{"domain": "myapp.com", "price": 17.29},
{"domain": "myapp.io", "price": 34.99}
],
"unavailable": ["myapp.ai"],
"summary": {
"cheapestAvailable": {"domain": "myapp.com", "price": 17.29},
"shortestAvailable": {"domain": "myapp.io", "price": 34.99}
}
}
check_handles(username, platforms?, only_report_available?)
Check social media handle availability across platforms.
Parameters:
| Name | Type | Default | Description |
|---|---|---|---|
username | str | required | The username/handle to check |
platforms | list[str] | all platforms | Platforms to check |
only_report_available | bool | false | If true, omit unavailable handles from response |
Supported platforms: instagram, twitter, reddit, youtube, tiktok, twitch, threads, bluesky, github, snapchat, pinterest, kick, substack
All platforms are checked in parallel; a typical call takes 2–5 seconds. Instagram, Threads and Reddit use a headless Chromium browser (installed automatically on first use).
How each platform is checked, and what "available" means:
| Platform | Method | "Available" means |
|---|---|---|
twitter | X signup's username availability endpoint | X would let you register it |
instagram | Profile page, then Instagram signup's username validation | Instagram would let you register it |
threads | Follows Instagram (Threads handles are Instagram usernames); profile page for the URL | Instagram would let you register it |
bluesky | com.atproto.temp.checkHandleAvailability on bsky.social, for <name>.bsky.social | Bluesky would let you register it |
reddit | Reddit's username_available endpoint | Reddit would let you register it |
tiktok | Profile page's embedded user status | No account found (banned or removed accounts look the same) |
youtube | @handle page status | No channel found (reserved or terminated handles may look the same) |
twitch | Twitch web GraphQL user lookup, including suspended and deleted accounts | No account found |
github | GitHub REST API user lookup (users and organizations; 60 lookups/hour without a token) | No account found (reserved names may look the same) |
snapchat | snapchat.com/@username page status | No account found (reserved names may look the same) |
pinterest | Profile page title, or the page's embedded "User not found" error | No account found (reserved names may look the same) |
kick | Kick public channel API, plus the hyphenated slug Kick gives some underscore usernames | No channel found (Kick does not publish length rules, so only characters are validated) |
substack | Whether <name>.substack.com serves or redirects (publication subdomains, not Substack user handles) | No publication or reserved subdomain found |
A platform is reported as available only when it positively says so. Block pages, rate limits and unexpected responses are reported as errors (in unavailable, with an error field), never as available or taken.
Response:
{
"available": ["instagram", "tiktok", "youtube"],
"unavailable": [
{"platform": "twitter", "url": "https://x.com/myapp"},
{"platform": "reddit", "url": "https://www.reddit.com/user/myapp"},
{"platform": "bluesky", "note": "Reserved or not allowed by Bluesky"},
{"platform": "twitch", "error": "Twitch returned HTTP 503"}
]
}
check_subreddits(names, only_report_available?)
Check subreddit name availability on Reddit.
Parameters:
| Name | Type | Default | Description |
|---|---|---|---|
names | list[str] | required | Subreddit names to check (with or without r/ prefix) |
only_report_available | bool | false | If true, omit unavailable subreddits from response |
Response:
{
"available": ["mynewsubreddit"],
"unavailable": [
{"name": "programming", "subscribers": 6835000},
{"name": "privatesubreddit", "note": "private"}
]
}
Subreddits are looked up through Reddit's about.json inside a headless browser session. note carries Reddit's reason when a subreddit is held but not public (private, gold_only, quarantined, banned, ...), or explains why a name is not a valid subreddit name.
check_everything(components, tlds?, platforms?, method?, require_all_tlds_available?, only_report_available?, also_include_hyphens?)
Comprehensive check across domains and social media. Generates name combinations from components, checks domains first (fast), then checks social handles for names that pass the domain check.
Parameters:
| Name | Type | Default | Description |
|---|---|---|---|
components | list[str] | required | Name components to combine (e.g., ["red", "sweater"]) |
tlds | list[str] | ["com", "net", "org", "io", "ai"] | TLDs to check |
platforms | list[str] | all platforms | Social platforms to check |
method | str | "auto" | "auto", "rdap", or "namesilo" |
require_all_tlds_available | bool | false | If true, name must be available in ALL TLDs to qualify for handle checking |
only_report_available | bool | false | If true, omit unavailable items from response |
also_include_hyphens | bool | false | If true, also check domains for hyphenated versions (handles are not checked for these) |
Name Generation:
From components ["red", "sweater"], generates:
- Single components:
red,sweater - Concatenations:
redsweater,sweaterred - With
also_include_hyphens:red-sweater,sweater-red(domains only)
Hyphenated versions can appear in available_domains and domain_successful_basenames, but are never checked for social handles, so they never appear in available_handles, unavailable_handles or fully_available. A hyphen typed into a component (e.g., ["red-sweater"]) is the caller's own spelling and is checked for handles like any other name.
Response:
{
"available_domains": [
{"domain": "redsweater.com", "price": 17.29},
{"domain": "redsweater.io", "price": 34.99}
],
"domain_successful_basenames": ["redsweater", "sweaterred"],
"available_handles": {
"redsweater": ["instagram", "twitter", "youtube"],
"sweaterred": ["instagram", "tiktok"]
},
"unavailable_handles": {
"redsweater": [{"platform": "reddit", "url": "..."}]
},
"summary": {
"fully_available": ["sweaterred"],
"cheapest_domain": {"domain": "redsweater.com", "price": 17.29}
}
}
The fully_available list contains names that are available on ALL checked platforms.
Domain Lookup Methods
| Method | Description | Pricing | Speed |
|---|---|---|---|
auto | Uses NameSilo if API key configured, otherwise RDAP | With NameSilo | Fast |
rdap | Direct registry queries via IANA bootstrap | No | Fast |
namesilo | NameSilo API (requires API key) | Yes | Fast |
RDAP Limitations
RDAP is free and requires no API key, but has some limitations:
- TLD coverage - Not all top level domains (TLDs) have RDAP servers. Queries for unsupported TLDs will fail. RDAP works for .com, .net, .org, .app, .ai and more. See deployment.rdap.org for an up-to-date list (look for 'Yes' in the 'RDAP' column).
- No pricing - RDAP only reports availability, not the cost to register the domain.
- False positives - A domain may appear available via RDAP but actually be reserved or considered "premium" by registrars, making it effectively unavailable or prohibitively expensive to purchase.
For reliable results with pricing, configure a NameSilo API key.
Configuration
API key storage:
- macOS: Keychain (as
internet-names-mcp.namesilo) - Linux:
~/.config/internet-names-mcp/config.json - Windows:
%APPDATA%/internet-names-mcp/config.json
API key lookup order (first match wins):
- macOS Keychain (on macOS only)
- Environment variable (
NAMESILO_API_KEY) - Config file (fallback)
Development
Local Setup
The devsetup.sh script handles virtual environment creation and dependency installation:
git clone https://github.com/drewster99/InternetNamesMCP.git InternetNamesMCP
cd InternetNamesMCP
source devsetup.sh # Creates .venv, activates it, installs dependencies
playwright install chromium # Required for Instagram, Threads and Reddit checks (auto-installed on first use otherwise)
Options:
source devsetup.sh- Set up environment (default)source devsetup.sh --clean- Delete venv and cachessource devsetup.sh --clean --setup- Clean rebuild
API Key Configuration (Development)
When running from source, use the module directly instead of uvx:
source devsetup.sh # Activate environment first
python -m internet_names_mcp --setup # Configure API keys interactively
python -m internet_names_mcp --show-config # Show current configuration and key source
python -m internet_names_mcp --version # Show version
Running Tests
source devsetup.sh # Activate environment first
# Main test suites
python test_server.py # Full test suite - offline validation + online API tests
python test_mcp_interface.py # Tests via MCP protocol (stdio transport)
python test_rdap_client.py # Async RDAP client, rate limiter, batch queries
# Comparison/diagnostic tests
python test_methods.py # Compare RDAP vs NameSilo results for discrepancies
python test_rdap.py # Quick RDAP-only domain check
| Test File | Description |
|---|---|
test_server.py | Main test suite covering all MCP tools, edge cases, and API calls |
test_mcp_interface.py | Tests the server through actual MCP protocol via stdio |
test_rdap_client.py | Tests async RDAP client, rate limiting, and batch queries |
test_methods.py | Compares RDAP vs NameSilo to detect availability discrepancies |
test_rdap.py | Simple RDAP-only test for quick domain availability checks |
Interactive Testing
Use MCP Inspector for interactive debugging with a web UI:
npx @modelcontextprotocol/inspector .venv/bin/python -m internet_names_mcp
Project Structure
├── src/internet_names_mcp/
│ ├── __init__.py # CLI entry point
│ ├── __main__.py # Module runner
│ ├── server.py # MCP server
│ ├── config.py # Configuration management
│ ├── rdap_bootstrap.py # RDAP bootstrap cache
│ └── rdap_client.py # Async RDAP client
├── pyproject.toml # Package configuration
└── README.md
Data Files
RDAP Bootstrap Cache - Maps TLDs to their authoritative RDAP servers (auto-downloaded from IANA):
- macOS/Linux:
~/.cache/internet-names-mcp/rdap_bootstrap.json - Windows:
%APPDATA%/internet-names-mcp/rdap_bootstrap.json
The cache is automatically refreshed when expired (default 24h TTL from IANA's Cache-Control headers).
Troubleshooting
"playwright not installed" or Chromium errors
Install Playwright browser:
playwright install chromium
Or with uvx:
uvx --from playwright install chromium
"Instagram is throttling username checks"
Instagram limits how many signup username validations one IP address can make. While throttled it answers "not available" for every name; the server detects this with a random control name and reports an error (noting whether a profile exists) instead of a wrong answer. In testing the throttle tripped after roughly 20–40 checks from one IP address and cleared within about 25 minutes. Taken names are settled from the profile page and do not count toward it.
Reddit errors ("bot-check page" or "without data")
Reddit blocks non-browser clients and rate-limits by IP address. The server passes Reddit's JavaScript check in a headless browser and keeps the resulting Reddit cookies in memory (never on disk) for the life of the server process, because clearing a fresh challenge on every call makes Reddit escalate to a CAPTCHA. If Reddit still asks for a CAPTCHA, the error says so; the server never attempts to solve it. Wait a while and retry.
Copyright
Copyright (C) 2026 Nuclear Cyborg Corp