aijobsboom-x402

x402-gated data API for the aijobsboom.com skilled-trades job board, plus an MCP server wrapper that pays for calls on an agent's behalf.

Live API: https://aijobsboom-x402.sonofgus.workers.dev

What this is

Rollup statistics from the aijobsboom pipeline — counts, not listings. Every pipeline run appends one row: total rows pulled, rows removed per scrub rule, per-trade counts (electrician, HVAC, lineman, welder, mechanic, CDL), and perk-flag counts (fair-chance, will-train, per diem, day-1 benefits, sign-on bonus, travel paid, employer-posted six-figure, HOT count, North Dakota, Montana). No job titles, employer names, posting URLs, or descriptions are served anywhere in this dataset — see methodology.md for why.

Related MCP server: GoldBean

Endpoints

Payment is enforced via x402 — no accounts, API keys, or subscriptions required.

MCP server

mcp-server.js wraps the API as two MCP tools, get_snapshot and get_series. It holds its own Base wallet (funded with USDC) and pays for each gated call itself via x402-fetch, so the calling agent just invokes a tool and gets JSON back.

npm install
AIJOBSBOOM_API_BASE=https://aijobsboom-x402.sonofgus.workers.dev \
X402_CLIENT_PRIVATE_KEY=0x... \
npm run mcp

X402_CLIENT_PRIVATE_KEY is the 0x-prefixed private key of a Base wallet funded with USDC — never commit this.

Worker deployment

npm install
wrangler secret put X402_WALLET      # the Base address that receives USDC
wrangler secret put FACILITATOR_URL  # optional, overrides the default facilitator
npm run deploy

Data updates: after each pipeline.py run in the parent repo, copy the fresh archive/stats.csv over data/stats.csv, run npm run teaser to refresh teaser.html, then redeploy. The CSV is bundled into the Worker at build time — there's no runtime data fetch.

Docs

Full column dictionary, scrub-rule writeup, and ToS constraints: methodology.md.