nanoparse-mcp
OfficialSummary: NanoParse MCP lets your agent fetch any web page and get back clean, structured Markdown plus metadata — no account or API key needed.
Fetch and convert web pages to Markdown — call
nanoparse_fetchwith anyurlto render a page (JavaScript/SPA included) and receive clean, readable Markdown instead of raw HTML.Get structured metadata alongside content — responses include title, author, dates, language, word count, OpenGraph and schema.org JSON-LD, so agents can cite and reason about the page.
Request extraction diagnostics on demand — set
debug: trueto receive extra detail on how content was isolated (contentSelector, metadataSource, cleaning stats, timing) for troubleshooting or tuning.Pay per parse with no signup — the schema advertises a free allowance, then $0.0175 USDC on Base via x402 micropayments; no account or API key is required.
(Note: the schema documents only nanoparse_fetch; the README also advertises a nanoparse_status quota/balance tool that isn't in this schema, and the two sources disagree on free-tier size — 50 free per IP here vs. 10/month signed-wallet or 3 anonymous — and on price — $0.0175 vs. $0.002.)
Allows fetching arXiv pages, such as paper abstract pages, and converting them into clean Markdown with metadata and Litmus trust signals for agent reading and citation.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@nanoparse-mcp fetch https://example.com/article as markdown"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
NanoParse MCP
The easiest way for AI agents to read the web. NanoParse MCP is a hosted Model Context Protocol server at https://nanoparse.app/mcp that turns any URL into clean Markdown your agent can actually use — with machine-readable trust signals attached.
No install. No wallet key. No .env. Connect once and your agent can read the web.
Why agents choose NanoParse over Firecrawl or Jina Reader: every page comes back with Litmus — 20 machine-readable signals that tell your agent whether to trust and cite the page before it reads a word. Pay $0.002 per parse via x402 once the free tier runs out. Full JavaScript/SPA rendering, native GFM tables. No account, no subscription, no API key.
Quick start
Add this to your MCP client config:
{
"mcpServers": {
"nanoparse": {
"type": "http",
"url": "https://nanoparse.app/mcp"
}
}
}Restart your client, and your agent auto-discovers two tools:
nanoparse_fetch(url)— fetch any web page and get clean Markdown backnanoparse_status()— check free quota, wallet and USDC balance before spending
That's it. Your agent can now fetch pages from inside any MCP-compatible client (Claude, Cursor, Hermes, Windsurf, and others).
Related MCP server: mdapi-mcp
What you get
nanoparse_fetch renders the page in a real headless browser (JavaScript and all), isolates the actual article from nav, ads, and cookie banners, and returns:
Clean Markdown — GFM tables, callouts, math, and footnotes preserved
Metadata — title, author, dates, language, word count, OpenGraph, schema.org JSON-LD
Litmus — machine-readable signals your agent can reason over: source authority, freshness, structural trust, content density, syndication and paywall detection
nanoparse_status tells your agent exactly where it stands: free parses remaining, wallet and USDC balance, and whether the next call will succeed without payment. Agents call it before spending.
Real output
Every parse returns markdown + metadata + litmus. Here is the full litmus object from a real parse of an arXiv paper (examples/example-output.json — markdown truncated there, litmus complete and unmodified):
{
"litmus": {
"source": { "type": "academic", "authority": 0.85, "rationale": "academic source" },
"freshness": { "published": null, "age": "unknown", "stale_warning": "No publish date found — verify independently", "modified": null, "has_correction": false, "correction_note": null },
"tldr": "Abstract page for arXiv paper 2401.14295: Demystifying Chains, Trees, and Graphs of Thoughts",
"key_figures": [ { "value": "2,247", "context": "Thu, 25 Jan 2024 16:34:00 UTC (2,247 KB)" } ],
"structural_trust_score": { "level": "medium", "trust_factors": ["High-authority source", "No publish date found — verify independently"] },
"hedge_language": { "density": "low", "markers_found": 0, "per_500_words": 0 },
"content_density": { "thin_content": false, "substantive_sentence_ratio": 0.76 },
"syndication": { "detected": false, "note": null },
"paywall": { "detected": false, "note": null },
"content_type": "academic_paper",
"reading_time": { "minutes": 4 },
"is_archived": { "archive_org": true, "archive_is": false },
"outbound_authority": { "score": 0.54, "top_domains": ["ui.adsabs.harvard.edu", "scholar.google.com"] },
"boilerplate_ratio": { "ratio": 0.08, "confidence": "high" }
}
}The agent learns before reading a word: this is an academic source (0.85 authority), structural trust is medium — high authority, but no publish date found, so verify independently. It also gets the one-sentence summary, the key numbers, and confirmation there's no paywall or syndication flag. See examples/agent-conversation.md for a full worked round-trip.
Payments
10 free parses per month with a signed wallet (sign the claim message — no payment), or 3 anonymous free tries per network. No account, no API key, no human required.
After that: $0.002 per parse (flat, no subscriptions) via x402 — an open micropayment protocol. Your agent's wallet signs the payment automatically. To give your agent a Coinbase wallet, run
npx @coinbase/payments-mcpand connect it alongside NanoParse.
Copy-paste recipe for both paths — the wallet-MCP one with no code, and a ~40-line REST client — including the exact 402 → sign → retry loop and the error codes worth branching on: examples/pay-with-x402.md.
How payments work under the hood
When an agent without free quota calls nanoparse_fetch, the endpoint responds with an HTTP 402 — Payment Required carrying x402 payment instructions. The agent's wallet signs a $0.002 USDC (Base) transfer, the facilitator settles it, and the parse proceeds. The hosted endpoint handles the entire challenge → settlement flow — the agent never needs a private key on your machine.
Pricing vs Firecrawl
NanoParse | Firecrawl Hobby | |
Billing | $0.002 per parse, flat | $16/mo subscription, credits expire monthly |
5,000 parses | $25.00 | $16 base — but ~$145 with JSON/extraction modes (9× credit burn) |
Account / API key | None | Required |
Agent can pay directly (x402) | Yes | No |
Full math, including when Firecrawl's base rate wins at scale: examples/costs.md.
Example
You: Fetch https://arxiv.org/abs/2401.14295 and summarize the argument.
Agent: [calls nanoparse_fetch(url)]
→ clean Markdown + Litmus (academic, authority 0.85, 4-min read,
no date found — verify, no paywall)
→ summarizes from primary content, cites authority and freshnessRepository note
The nanoparse-mcp npm package and the local stdio client were the
first-generation integration, which held a wallet private key
(NANOPARSE_WALLET_KEY) and signed x402 payments on your machine. Both are
retired: the hosted endpoint above now handles the full payment flow
itself, so no local client or wallet key is needed. The retired client code
is preserved in archive/v1-local-client/ for
history, and the npm package is deprecated. Use the hosted endpoint — it
is the current, supported integration.
Docs
Quick start & integration: nanoparse.app/quickstart
Compare & pricing: nanoparse.app/compare
API + MCP endpoint: nanoparse.app
Security policy: SECURITY.md — report a vulnerability to hello@nanoparse.app (also published at
https://nanoparse.app/.well-known/security.txt)
MIT licensed.
Available Tools
1 toolnanoparse_fetchAInspect
Fetch a web page and return clean, structured Markdown with metadata. First 50 parses free per IP. After that, pay $0.0175 USDC on Base via x402. No account or API key required.
Inputs: url (required), debug.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 1 tool
Only one tool exists, so there is no possibility of confusion or overlap. The single tool has a clearly defined purpose of fetching and converting web pages to markdown.
The tool name follows a consistent verb_noun pattern: 'nanoparse_fetch'. This is clear and predictable, though with only one tool, consistency is trivial.
Having only one tool feels thin for a server, even for a focused purpose. While the tool covers a specific function, agents might expect additional related operations or options.
For the stated purpose of fetching a web page and returning structured markdown with metadata, the single tool provides complete coverage. No other operations are necessary to fulfill the server's promise.
Related MCP Connectors
Programmatic web-scraping MCP server powered by x402 micro-transactions on Base.
Monetize any MCP server: x402 paywall, pay-per-call billing in USDC on Base, agent marketplace.
Pay-per-use MCP tools for AI agents. USDC on Base via x402. No API keys.
Pay-per-fetch web extraction for agents: POST a URL, get clean markdown. x402 USDC on Base.
Related MCP Servers
- AlicenseAqualityBmaintenanceAn MCP server that enables AI agents to access paid AI inference and web tools via HTTP 402 micropayments in USDC on Base, using the agent's wallet as identity.36122 npmMIT
- AlicenseNot gradedqualityDmaintenanceMCP server for converting HTML to Markdown and extracting page data, with pay-per-call using USDC on Base.MIT

fetcher.cloudofficial
AlicenseNot gradedqualityDmaintenanceMCP server that gives AI agents web capabilities such as rendering pages, extracting content, resolving links, and inspecting domains, with per-call USDC payments via x402.MIT- AlicenseNot gradedqualityDmaintenanceMCP server providing 11 pay-per-call web intelligence tools (page reading, PDF extraction, RSS parsing, screenshots, summarization, structured data extraction) for AI agents, paid automatically in USDC via the x402 protocol on Base mainnet with no API keys required.38 npmMIT