Clearspar Part 135 MCP
An MCP server that turns Claude into a US Part 135 market analyst over live FAA operator data, computing concentration ratios, HHI, fleet distributions, and rankings.
README
Clearspar Part 135 MCP
An MCP server that turns Claude into a US Part 135 market analyst over my live Clearspar FAA operator dataset — 1,890 active Part 135 charter certificate holders scraped from the official FAA xlsx (fleet size derived from tail counts, state-mapped).
It wraps my real, key-gated endpoint GET /api/prospects on clearspar.binnacleai.com. The tools compute — concentration ratios, HHI, fleet distributions, rankings — they don't just dump rows. Ask "How concentrated is Alaska's Part 135 charter market?" and Claude pulls all 303 Alaska operators live and answers with real math.
This is a thin analyst layer over my own production data — not a public FAA API. It's read-only and key-gated.
What it is (honest scope)
The dataset reliably populates three dimensions: operator name, state (2-letter USPS), and fleetSize (FAA tail count), plus a crmStatus pipeline field. Every tool here is built on those. Sparse/empty fields (email, phone, website, style, certificate dates, aircraft type, lat/lng) are deliberately not exposed as tools — see "Not built" below — so the server never fabricates a metric it can't back with data.
Tools
| Tool | What it computes |
|---|---|
market_summary(state, topN?) |
Operator count, total tails, avg & median fleet, largest operator, top-5 tail share, HHI concentration, top operators. |
fleet_ranking({state?, limit?}) |
Operators ranked by fleet size. State-scoped or the full ~1,890 universe. |
compare_operators(names[]) |
Side-by-side of 2+ operators (state, fleet, certNo, CRM status); who's largest, shared state. |
competitive_landscape(state) |
Fleet-size distribution buckets, top-1/3/5 concentration, HHI with plain-English read, market leaders. |
crm_pipeline_summary({state?}) |
Live counts across CRM stages (NEW/CONTACTED/REPLIED/DEMO/CLOSED/DEAD). |
Not built (data doesn't support it)
- Proximity / radius / "operators near X" — there is no lat/lng, base-airport, or ICAO field. Geography is state-level only.
- Operator-type segmentation (bush vs scheduled vs tour) — the
stylefield exists but isNULLin 100% of records. - Contact / email / phone outreach tools — email, phone, and website are populated on ~0.5% of records (known enrichment gap).
- Certificate-age / aircraft-make-model tools — no cert-issue-date or aircraft-type field exists; only a derived
fleetSizeinteger.
Install
cd clearspar-part135-mcp
npm install
npm run build
Configure
Copy .env.example and fill in the key:
cp .env.example .env
CLEARSPAR_BASE_URL=https://clearspar.binnacleai.com
CLEARSPAR_API_KEY=<the DIGEST_API_KEY value>
The endpoint is key-gated (returns 401 without a valid key — that gate is verified). The key is the DIGEST_API_KEY env var on the takeoff-app Docker container (host 144.202.116.229):
# retrieve without printing it into history/logs
docker exec takeoff-app printenv DIGEST_API_KEY
CLEARSPAR_API_KEY is required — the server refuses to call the API without it.
Claude Desktop config
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"clearspar-part135": {
"command": "node",
"args": ["/Users/johnthomasair/code/clearspar-part135-mcp/build/index.js"],
"env": {
"CLEARSPAR_BASE_URL": "https://clearspar.binnacleai.com",
"CLEARSPAR_API_KEY": "<the DIGEST_API_KEY value>"
}
}
}
}
Restart Claude Desktop. The five tools appear under the plug icon.
Demo prompts
- "How concentrated is Alaska's Part 135 charter market?"
- "Give me a market summary for Hawaii's Part 135 operators."
- "Rank the top 15 Part 135 operators nationwide by fleet size."
- "Compare Grant Aviation and Bering Air."
- "What's the competitive landscape in Texas — is it fragmented or dominated by a few?"
- "Where does my CRM pipeline stand for Alaska?"
How it talks to the API
GET /api/prospects?state=&crmStatus=&limit=&offset=with headerX-API-Key.- Response:
{ "total": <count of matching ACTIVE rows>, "rows": [ <Prospect> ] }. The server always filtersstatus: 'ACTIVE'. limitis hard-capped at 500 server-side. Any tool needing full coverage paginates byoffsetuntil it has alltotalrows, and caches the ~1,890-row universe for the process lifetime (the data only changes when the FAA scraper re-runs).- There is no free-text search param, so
compare_operatorsmatches names client-side. - Read-only: only
GETis ever called (the route also hasPOST/PATCH, which this server never uses).
Live data sanity (as of build)
total = 1890 active operators · AK = 303 · HI = 28 · all certType = "135" · all crmStatus = "NEW". Verified against the live endpoint; fleetSize ranges 1–386 nationwide (1–42 in AK).
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.
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.
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.
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.