Back to Browse

Doc Extract Api MCP Server

Developer ToolsUse Caution4.2MCP RegistryLocal
Free

Server data from the Official MCP Registry

SchemaLock — paid extraction: JSON from docs via your schema, x402/USDC.

About

SchemaLock — paid extraction: JSON from docs via your schema, x402/USDC.

Security Report

4.2
Use Caution4.2High Risk

This is a legitimate paid document extraction service built on x402 (USDC micropayments). The codebase is well-structured with proper authentication via cryptographic payment signatures, appropriate input validation, and clear separation of concerns. However, the MCP server requires users to supply their own funded wallet private keys and makes real financial payments on every tool invocation, which creates moderate financial and user-experience risks that should be clearly documented. The core extraction API itself is sound, but the MCP distribution model warrants careful user communication about costs and wallet management. Supply chain analysis found 10 known vulnerabilities in dependencies (0 critical, 4 high severity). Package verification found 1 issue.

7 files analyzed · 18 issues found

Security scores are indicators to help you make informed decisions, not guarantees. Always review permissions before connecting any MCP server.

Permissions Required

This plugin requests these system permissions. Most are normal for its category.

HTTP Network Access

Connects to external APIs or services over the internet.

env_vars

Check that this permission is expected for this type of plugin.

process_spawn

Check that this permission is expected for this type of plugin.

What You'll Need

Set these up before or after installing:

EVM private key for a wallet funded with real USDC on Base. Every tool call pays the live x402 API from this wallet — there is no free extraction path.Required

Environment variable: WALLET_PRIVATE_KEY

How to Install

Add this to your MCP configuration file:

{
  "mcpServers": {
    "io-github-thestarboy9696-doc-extract-api": {
      "env": {
        "WALLET_PRIVATE_KEY": "your-wallet-private-key-here"
      },
      "args": [
        "-y",
        "doc-extract-api"
      ],
      "command": "npx"
    }
  }
}

Documentation

View on GitHub

From the project's GitHub README.

doc-extract-api

Document-to-structured-JSON extraction API. Takes a PDF/PNG/JPEG and returns clean JSON. Four fixed document types (invoice, receipt, contract, resume) get a matched schema plus domain-specific validation; /extract/custom accepts any JSON Schema, so a caller isn't limited to those four types. Built for autonomous AI agents: no signup, no API key, no human checkout — an agent discovers the price and pays per call via x402 v2 (USDC), entirely over HTTP, and can find this service programmatically via the x402 Bazaar discovery catalog. Also published as an MCP server (doc-extract-api-mcp on npm, and on the official MCP Registry) — but that's a client for the same paid API, not a free alternative: every MCP tool call is a real x402 payment from a wallet you supply, not a locally-run, unmetered extraction. There is no free path to this service.

Why x402, not Stripe

Stripe Checkout is built for humans — it collects a card through a UI, and its payment button actively resists non-trusted, programmatic (agent-driven) submission. x402 revives the HTTP 402 Payment Required status code: hit a paid endpoint with no payment, get back a 402 whose PAYMENT-REQUIRED header carries a complete, machine-readable manifest (price, network, receiving address, asset, and — via the Bazaar extension — a full input/output schema with a real example). Retry with an X-PAYMENT header carrying a signed USDC transfer authorization, and the request goes through. No account, no key, no human in the loop — the discovery step and the payment step are the same HTTP round trip.

Payment is only settled after the extraction succeeds — @x402/hono's middleware verifies the payment signature before running the handler, but only executes the actual on-chain transfer if the handler returns a non-error status (confirmed: this project's paymentMiddlewareFromConfig call skips settlement whenever the wrapped route responds ≥400). A failed extraction costs the caller nothing.

Discovery: the x402 Bazaar

The Bazaar (GET {facilitator}/discovery/resources) is a live, queryable catalog of x402 services — ~15k listings as of this writing, searchable by agent clients (client.extensions.bazaar.search({ query: "..." })). Getting listed isn't a submission form: a facilitator catalogs a resource automatically the first time it processes a real payment that echoes the route's declared bazaar extension. src/lib/discovery.ts declares that extension for all four routes — full input schema (multipart form-data, a file field) and a real, verified output example — so once a route's first live mainnet payment lands, it becomes findable in the catalog with enough detail for an agent to integrate correctly without reading any docs. As of this writing the /extract/invoice catalog entry is still showing a stale pre-migration record rather than this rich content — see git history / ask for status; /extract/receipt, /extract/contract, and /extract/resume haven't had a mainnet payment yet so aren't catalogued at all.

Positioning (from surveying the live Bazaar before writing the listing copy): generic OCR competitors are free but return raw unvalidated text; generic structured-extraction competitors charge ~$0.10/call and require the caller to already have plain text, not a raw file. This is the only listing found doing both — raw PDF/image in, a fixed validated schema out. Priced at $0.05/call (still half the nearest real competitor). See src/lib/discovery.ts for the exact listing copy and each route's tags (serviceName/tags/description).

Architecture

  • schemas/ — JSON Schema for invoice, receipt, contract, and resume extractions (also the tool input_schema sent to Claude, so the model is forced into this exact shape).
  • src/lib/claude.ts — calls the Anthropic Messages API with the document as a native document/image content block and a forced tool_choice, so the model can't return anything but the schema.
  • src/lib/validate.ts — post-extraction sanity checks: line items sum to subtotal, subtotal + tax (+ tip) ≈ total, dates are valid and not in the future, currency is a real ISO 4217 code, low-confidence fields get flagged.
  • src/lib/extract.ts — orchestrates one extraction pass, validates, and — if validation fails — retries once with a corrective prompt describing exactly what didn't add up. Never fails silently; returns validation_warnings alongside the data either way.
  • src/lib/customSchema.ts / src/routes/customExtractHandler.ts/extract/custom: same Claude forced-tool-use engine, but the tool's input_schema is a JSON Schema the caller supplies at request time instead of one of the four fixed schemas. There's no domain-specific validator (no fixed domain to check against), so this returns schema-conformant output, not fact-checked output — validation_warnings is always []. The caller's schema is bounded (≤6000 chars, depth ≤4, ≤30 properties per level, no $ref) and rejected with a 400 — which costs nothing, since x402 only settles on a non-error response — before any model call. Accepts either a file (binary PDF/image, same as the fixed routes) or content (raw text/HTML the caller already has — e.g. an already-scraped web page, ≤100,000 chars) — exactly one of the two, not both. The text/HTML path is labeled explicitly as data in the prompt sent to Claude, not instructions, since it's arbitrary caller-supplied content and could otherwise be a prompt-injection vector.
  • src/lib/facilitator.ts — picks the free default facilitator on testnet, or, on mainnet (X402_NETWORK = "base"), Coinbase's authenticated CDP facilitator plus Mogami and PayAI for additional Bazaar-equivalent discovery surface (CDP stays first/primary — earlier facilitators get precedence for verify/settle); maps this project's human-readable network names to the CAIP-2 identifiers x402 v2 uses (eip155:8453 etc). Important: src/index.ts caches the built payment middleware at module scope rather than rebuilding it per-request — with multiple facilitators, an unhealthy one adds its full timeout to initialize(), and rebuilding per-request meant paying that cost on every request (this broke the API entirely the first time multi-facilitator was tried). Caching means only a cold isolate's first request pays that cost.
  • src/lib/discovery.ts — the Bazaar listing content: descriptions, tags, and discovery extensions (schema + real example) for all four routes.
  • src/index.ts — Hono app on Cloudflare Workers. @x402/hono's paymentMiddlewareFromConfig guards /extract/*, built per-request from Env bindings (Workers only exposes env vars inside a request, not at module load time).
  • src/routes/extractHandler.ts — pure extraction logic; payment has already been verified by the time this runs.
  • mcp/server.ts — stdio MCP server exposing extract_invoice, extract_receipt, extract_contract, extract_resume, and extract_custom tools. Each tool call signs and sends a real x402 payment (via @x402/fetch's wrapFetchWithPayment, same mechanism as scripts/test-payment.mjs) from a wallet supplied via WALLET_PRIVATE_KEY, then calls the live HTTP API over the network — it does not run extraction locally or call Claude directly. No free path: a missing or unfunded wallet fails the same way it would calling the API directly.

One deliberate deviation from the original spec

The spec assumed a PDF→image conversion step (pdf2image/poppler). Cloudflare Workers can't run native binaries, and Claude's Messages API accepts PDFs natively as a document content block — so PDFs are sent to Claude as-is. This removes a whole pipeline stage and a dependency that wouldn't run on Workers anyway.

Setup

cd "doc-extract-api"
npm install

1. Secrets

Local dev — copy .dev.vars.example to .dev.vars and fill in a real key (gitignored):

cp .dev.vars.example .dev.vars

Production:

npx wrangler secret put ANTHROPIC_API_KEY

2. Wallet — where payments land

Edit wrangler.toml: X402_PAY_TO_ADDRESS must be a real wallet address you control before this can receive actual payments. It's a public receiving address (safe to commit, like a bank account number) — never put a private key anywhere in this project.

X402_NETWORK defaults to base-sepolia (testnet, free fake USDC from a faucet — good for verifying the whole flow with zero real money at risk).

X402_PRICE_PER_CALL defaults to $0.05, applied to all four routes.

2b. Going to mainnet (X402_NETWORK = "base")

The free default facilitator (x402.org/facilitator, used automatically on testnet) only supports testnet — mainnet requires an authenticated facilitator. This project is wired for Coinbase's CDP facilitator (src/lib/facilitator.ts):

  1. Create a CDP Portal account and a Secret API Key under Project → API Keys.
  2. npx wrangler secret put CDP_API_KEY_ID
  3. npx wrangler secret put CDP_API_KEY_SECRET
  4. Set X402_NETWORK = "base" in wrangler.toml and redeploy.

Pricing: free for the first ~1,000 settled transactions/month, then $0.001 each — payment verification itself is always free. If X402_NETWORK is "base" and these secrets aren't set, requests will fail loudly at the facilitator-config step rather than silently misconfiguring payments.

Getting listed in the Bazaar requires at least one real mainnet payment to go through — the facilitator catalogs the resource the first time it processes a payment that echoes the bazaar extension. Run scripts/test-payment.mjs with a real funded wallet once to trigger it (see below), then confirm with:

curl -s "https://api.cdp.coinbase.com/platform/v2/x402/discovery/resources?limit=200" \
  | grep -o 'doc-extract-api'

3. End-to-end payment test (scripts/test-payment.mjs)

Simulates a real autonomous agent: generates/uses a throwaway wallet, pays via x402 v2, and calls the live API. Fund a test wallet first — testnet USDC from a faucet (e.g. faucet.circle.com, Base Sepolia) for a free run, or a small amount of real USDC on Base if X402_NETWORK is "base":

node scripts/gen-test-wallet.mjs        # prints a fresh throwaway private key + address
TEST_WALLET_PRIVATE_KEY=0x... node scripts/test-payment.mjs

Prints the settlement details and the recipient's before/after USDC balance — confirmed working end to end on both Base Sepolia (testnet) and Base mainnet with real USDC.

4. Run locally

npm run dev
curl -X POST http://localhost:8787/extract/invoice -F "file=@/path/to/invoice.pdf"
# → 402 Payment Required; full payment + discovery manifest in the PAYMENT-REQUIRED header

/extract/custom takes a schema field (a JSON Schema string, type: "object"), an optional instructions field, and exactly one of file (binary) or content (raw text/HTML):

curl -X POST http://localhost:8787/extract/custom \
  -F "file=@/path/to/doc.pdf" \
  -F 'schema={"type":"object","properties":{"order_id":{"type":"string"},"ship_date":{"type":"string"}}}' \
  -F "instructions=Dates in YYYY-MM-DD"

Or from raw text/HTML you already have — no file needed:

curl -X POST http://localhost:8787/extract/custom \
  -F "content=<html><body><h1>Acme Corp</h1><p>Founded 1998, HQ in Austin, TX</p></body></html>" \
  -F 'schema={"type":"object","properties":{"company":{"type":"string"},"founded":{"type":"number"},"hq":{"type":"string"}}}'

An actual caller uses an x402-aware HTTP client (e.g. @x402/fetch on the agent's side) that handles the 402 → sign payment → retry loop automatically.

5. Deploy

npm run deploy

6. MCP server (paid, agent use)

export WALLET_PRIVATE_KEY=0x...   # a wallet funded with real USDC on Base — every call spends from it
npm run mcp

Or, once published, an MCP client can run it directly without cloning this repo:

npx doc-extract-api-mcp

Point an MCP-capable client at this stdio server to expose extract_invoice, extract_receipt, extract_contract, extract_resume, and extract_custom as tools. Every call pays the live API over the network via x402 — there is no local/free extraction path, and no ANTHROPIC_API_KEY is needed on the client side at all (the server holds that; callers only need a funded wallet).

Not yet done

  • Resumes carry real PII (name, contact info, work history). Nothing is persisted after the response is returned (same as every other document type here), but that's worth restating given the sensitivity — see src/lib/discovery.ts's resume description.

Reviews

No reviews yet

Be the first to review this server!