ProofFetch — Cited Web Evidence MCP server
online · 100% uptimeWeb pages your agent can quote: 1,000 full fetches a month free, no card, then $0.01 a page.
not scanned yet
0 stars
Reviews
Write oneNobody has reviewed ProofFetch — Cited Web Evidence yet.
If you have run it, two minutes of your experience saves the next person an afternoon.
ProofFetch — Cited Web Evidence tools
Tool list not cached yet. `describe` through the gateway fetches it live.
How ProofFetch — Cited Web Evidence compares
Against every graded AI and agents MCP server, from our own scans and uptime checks. Updated after each scan run.
- Safety score
- 88/100
- higher than 80% of 4,461 AI and agents servers
- Uptime
- 100%
- of our recent checks; the typical remote AI and agents server answers 100%
- Response time
- 394 ms
- faster than 38% of 2,334 AI and agents servers
Release history: 3 versions in the registry since 2026-09-06, latest 2026-10-29, typically one every 38 days.
Install ProofFetch — Cited Web Evidence in Claude Code, Cursor or VS Code
claude mcp add --transport http proof-fetch https://prooffetch.steellantern.com/mcp
What the publisher says
From the ProofFetch — Cited Web Evidence repository's README, as published. We do not edit it. Read it on GitHub
ProofFetch: cited web evidence for agents
Free check with a text preview (about the first 1,000 characters; short pages come back whole). Full text costs 1¢ a page either way: 500 pages for $5 by card, or $0.50 for 50 if your agent pays automatically through Stripe (the page it asked for, plus a key for 49 more).
Try the preflight · Current offer · OpenAPI · Support
For agents: llms.txt · A2A agent card · connection guide · MCP registry name io.github.equinoxaifinance-rgb/proof-fetch
Before paying: terms · privacy · refunds
This repository contains documentation and small client examples only. It does not contain the extraction engine, a payment credential, or a way to obtain the paid paragraphs free.
Start here
- Inspect a URL before giving it to an agent.
- Handle a payment challenge without accidental retries.
- With Node 22 or later, run node examples/preflight.mjs https://example.com from this repository. The example makes no payment.
- In n8n, inspect free metadata and stop before payment using the bounded client workflow.
This repository's documentation, discovery metadata and examples, including the n8n wrapper and guide in guides/n8n, are MIT licensed under the root LICENSE. The MIT licence does not cover the ProofFetch extraction engine or the hosted service at prooffetch.tryneverclosed.com: their source is not in this repository and is not licensed or distributed here. Paid paragraphs, upstream material, credentials and account or payment rights keep their own terms.
Agent skill
skills/cited-web-evidence/SKILL.md teaches an agent when and how to use ProofFetch: preflight first, pay only with the user's approval, and cite paragraph IDs. Install it into Claude Code, Cursor, Codex or another skills-aware agent with npx skills add equinoxaifinance-rgb/proof-fetch-docs --skill cited-web-evidence.
Connect an MCP client
Use Streamable HTTP at https://prooffetch.tryneverclosed.com/mcp (the original https://proof-fetch-agent-api.neoaethel.workers.dev/mcp keeps working). In Claude Code: claude mcp add --transport http proof-fetch https://prooffetch.tryneverclosed.com/mcp. Its free tools are prooffetchpreflight and prooffetchoffer. It also lists prooffetchevidence, which returns the paid text and needs a ProofFetch card-pack key sent as Authorization: Bearer . MCP Registry manifest: server.json (io.github.equinoxaifinance-rgb/proof-fetch). The free tools never return the paid page text. Clients paying through Stripe MPP use POST /v1/evidence; each $0.50 payment returns the page plus callBundle.apiKey for 49 more.
What you receive
Free preflight: fetch metadata, document statistics, content/raw hashes and deterministic prompt-injection indicators. Paid fulfillment adds paragraph IDs, text and a payment receipt. A hash identifies bytes; it does not certify that a source is true. An unflagged page is not a safety guarantee.
Supported: public HTML, XHTML, text and JSON, at most 1 MiB and five redirects. No login, cookies, caller headers, browser rendering or private-network destinations. Inspect the live offer for current limits before integration.
Design-partner invitation
Shortened. The full README is on GitHub.
Nothing above is checked by us. What we check is on the safety report.
Grade history
- 2026-10-16restoreC → Ascore 88
ProofFetch — Cited Web Evidence: common questions
- Is ProofFetch — Cited Web Evidence MCP server safe?
- Yes, by our scan: it is graded A (88/100). Read the ProofFetch — Cited Web Evidence safety report
- How do I install ProofFetch — Cited Web Evidence?
- It runs remotely at prooffetch.steellantern.com. Add it to Claude Code, Cursor, VS Code or Claude Desktop with the snippets above, or call it through the mcp.market gateway without installing anything.
- Does ProofFetch — Cited Web Evidence need an API key?
- Not as far as the registry entry and our scan can tell: no credentials are declared or required.
- Is ProofFetch — Cited Web Evidence maintained?
- The last commit was 21 days ago (2026-10-15). The latest release is v0.3.1.
- Is ProofFetch — Cited Web Evidence up?
- 100% of our last 28 checks got an answer. We check remote servers about four times a day.
- What can I use instead of ProofFetch — Cited Web Evidence?
- Servers from other publishers that do the same job: Robin: AI-Powered Dark Web OSINT MCP server, Charlotte MCP server and agent-device MCP server.
- How does ProofFetch — Cited Web Evidence compare with other AI and agents MCP servers?
- safety score 88/100 (higher than 80% of 4,461 AI and agents servers); uptime 100% (of our recent checks; the typical remote AI and agents server answers 100%); response time 394 ms (faster than 38% of 2,334 AI and agents servers). Full comparison.
Alternatives to ProofFetch — Cited Web Evidence
Same job from other publishers: the closest match first, then the best rated.
- Robin: AI-Powered Dark Web OSINTDark web OSINT over Tor: search onion engines, scrape pages, report with your own model.not reviewedEstablishedA
- CharlotteRenders web pages into structured, agent-readable representations using headless Chromium.not reviewedEstablishedA
- agent-deviceOpen source device layer for AI agents: control mobile, web, TV, and desktop apps over MCPnot reviewedWidely usedA
- HOL GuardLocal-first AI agent security evidence and approval workflows through HOL Guard's stdio MCP server.not reviewedEstablishedA
- Watch SkillWatch video and live sessions, keep timestamped evidence, and verify an agent's own work.not reviewedEstablishedB