SmartFetch
Click on "Install 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., "@SmartFetchFetch https://example.com and summarize the article"
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.
SmartFetch V1.8 — MCP discoverability
SmartFetch takes a public web URL and returns clean agent-ready text, Markdown, links, metadata, and the retrieval method used. It tries cheap HTTP retrieval first and falls back to a real Chromium browser when needed.
What changed in V1.8
The paid MCP
fetch_webpagetool now declares official x402 Bazaar metadata for agent discovery without changing its execution or payment behavior.The root
server.jsonadvertises the existing/mcpStreamable HTTP endpoint as a remote-only server for the Official MCP Registry.No Registry publication, deployment, Railway variable change, or payment is performed by this repository update.
All V1.7 HTTP payment, Bazaar, and native MCP behavior remains unchanged:
Base Sepolia remains the default x402 network and continues to use
https://x402.org/facilitatorwithout CDP credentials.Base mainnet can use Coinbase's authenticated CDP facilitator when explicitly selected.
Mainnet startup requires both CDP API credentials and fails closed for invalid credentials or missing facilitator support.
/metadistinguishes enabled testnet and mainnet payment modes.cdp-sdk, FastAPI, Uvicorn, and x402 are exactly pinned so infrastructure upgrades are explicit.
The retrieval engine, extraction/browser behavior, SSRF protections, request limits, rate limiting, and concurrency controls are unchanged.
Related MCP server: SteadyFetch
API
GET /health
Returns service health/version.
GET /meta
Machine-readable service description.
POST /fetch
{
"url": "https://example.com/article",
"max_chars": 20000,
"force_browser": false
}max_chars is optional (minimum 1,000; default 20,000; maximum 50,000).
Example response fields:
{
"success": true,
"requested_url": "https://example.com/article",
"final_url": "https://example.com/article",
"status_code": 200,
"render_method": "http",
"title": "…",
"content": "…",
"markdown": "…",
"links": [],
"word_count": 1000,
"content_hash": "…",
"truncated": false,
"elapsed_ms": 350,
"request_id": "…",
"service_version": "1.8.0"
}When x402 payment protection is enabled, the PAYMENT-REQUIRED header for
POST /fetch includes a Bazaar declaration with the url, max_chars, and
force_browser input contract plus a representative successful response.
/, /health, and /meta remain free and do not advertise Bazaar metadata.
MCP
Remote MCP clients connect to /mcp using Streamable HTTP. The server exposes
exactly one tool:
fetch_webpage
url: required string
max_chars: optional integer (default 20000, minimum 1000, maximum 50000)
force_browser: optional boolean (default false)The tool reads, fetches, scrapes, or extracts public webpages for agents and
returns the existing SmartFetch text, Markdown, links, and retrieval metadata.
It uses the same SSRF validation, executor, request timeout, concurrency cap,
output cap, HTTP retrieval, and browser fallback as POST /fetch; it does not
make an HTTP request back to the public API.
When x402 is enabled, an unpaid tools/call returns the native MCP x402 payment
challenge. A valid payment is verified before retrieval and settled once after
successful tool execution. The network, public payee, and price come from
X402_NETWORK, X402_PAY_TO, and X402_PRICE. No seller private key or wallet
secret is accepted or required.
The unpaid MCP payment challenge includes a Bazaar declaration for
fetch_webpage with transport streamable-http, the existing input contract,
and a representative SmartFetch output. Its payment resource remains
mcp://tool/fetch_webpage. The existing HTTP Bazaar resource for POST /fetch
is separate and unchanged.
The root server.json describes the public remote endpoint at
https://smartfetch-production-ea53.up.railway.app/mcp. Registry publication is
a separate authenticated operator action; adding the manifest does not publish
or deploy the service.
Run locally
python -m pip install -r requirements.txt
python -m smartfetch.serverThen:
curl http://127.0.0.1:8787/health
curl -X POST http://127.0.0.1:8787/fetch \
-H 'Content-Type: application/json' \
-d '{"url":"https://example.com"}'Windows users can run setup_windows.bat, then start_windows.bat.
With X402_ENABLED unset or false, startup requires no payment settings and /fetch remains free for local and remote testing.
x402 payment modes
Payment remains disabled unless explicitly enabled. Base Sepolia is the default and requires only the public receiving address:
X402_ENABLED=true
X402_PAY_TO=0xYOUR_PUBLIC_RECEIVING_ADDRESS
X402_PRICE=$0.005
X402_NETWORK=eip155:84532X402_PRICE and X402_NETWORK use the values shown when omitted. The default price remains $0.005.
Base mainnet must be selected explicitly and uses Coinbase's authenticated CDP facilitator:
X402_ENABLED=true
X402_PAY_TO=0xYOUR_PUBLIC_RECEIVING_ADDRESS
X402_PRICE=$0.005
X402_NETWORK=eip155:8453
CDP_API_KEY_ID=YOUR_CDP_API_KEY_ID
CDP_API_KEY_SECRET=YOUR_CDP_API_KEY_SECRETMainnet requires both CDP credentials. Missing, invalid, or unusable credentials, unsupported Base-mainnet exact payments, or any facilitator/middleware initialization failure aborts startup. SmartFetch never falls back to free access or the testnet facilitator when mainnet is selected. Every network other than Base Sepolia and Base mainnet is rejected.
Only POST /fetch is protected. /health, /, and /meta always remain free. Enabled /meta responses report x402-enabled-testnet or x402-enabled-mainnet without exposing configuration values.
SmartFetch uses X402_PAY_TO as the public receiving address and does not need CDP_WALLET_SECRET. Never provide a seller MetaMask private key or recovery phrase: SmartFetch does not require, read, accept, log, or store either one.
Local validation
python tests/security_smoke.py
python tests/api_local_smoke.pyapi_local_smoke.py intentionally enables private-network access only inside the test process so it can use a localhost fixture.
Container
docker build -t smartfetch:v1.8 .
docker run --rm -p 8787:8787 smartfetch:v1.8The container installs Chromium automatically.
Public deployment
This package is Docker-ready. Railway can deploy the included Dockerfile and railway.json; any Docker host that provides outbound HTTPS and enough RAM for one Chromium process should also work.
Recommended initial environment variables:
MAX_CONCURRENT_BROWSERS=1
MAX_CONCURRENT_FETCHES=4
DEFAULT_MAX_OUTPUT_CHARS=20000
MAX_OUTPUT_CHARS=50000
FETCH_TIMEOUT_SECONDS=12
BROWSER_TIMEOUT_SECONDS=15
TOTAL_REQUEST_TIMEOUT_SECONDS=25
RATE_LIMIT_PER_MINUTE=30
RATE_LIMIT_BURST=10
X402_ENABLED=falseDo not set ALLOW_PRIVATE_NETWORK=1 in production.
Important SSRF deployment note
Application-level URL validation is included, but browser rendering can execute page subresources. For a paid public deployment, the browser container should additionally be isolated by host/network egress policy from cloud metadata and private RFC1918 networks. We will harden this again before removing the temporary rate limit and before x402 launch.
Test the public deployment
On Windows run test_remote_windows.bat, paste the public URL, and upload tests/remote_20_results.json.
Or:
python tests/remote_20.py https://YOUR-PUBLIC-URLOur deployment gate remains 18/20 minimum, including all five forced-browser requests.
Payment launch status
V1.8 preserves the existing Base Sepolia and Base mainnet payment configuration. This repository change does not itself publish to the MCP Registry, deploy the service, or modify Railway variables. The active payment network is controlled by X402_NETWORK: eip155:84532 for Base Sepolia or eip155:8453 for Base mainnet.
This server cannot be installed
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseAqualityBmaintenanceEnables AI agents to read web pages reliably, returning clean markdown content, hyperlinks, and metadata without navigation or ad noise.314MIT
- AlicenseNot gradedqualityDmaintenanceProvides AI agents with reliable web fetching capabilities, handling retries, caching, and anti-bot bypass automatically.MIT
- AlicenseAqualityCmaintenanceEnables AI agents to fetch any web page as clean markdown or screenshot it, turning URLs into LLM-ready context.211MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to fetch and extract clean, readable content from web pages, and search within pages for specific queries, without needing a full browser.1MIT
Related MCP Connectors
Web intelligence for AI agents: fetch, render, extract, research. x402 micropayments, no API keys.
The web capability layer for AI agents: render, extract, DNS, SSL, WHOIS & more via x402.
Read a URL as clean markdown, screenshot a website, url to PDF. Web access for agents, no signup.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Friezaaaa/smartfetch'
If you have feedback or need assistance with the MCP directory API, please join our Discord server