What changed
FACT (per signals): the MCP spec cuts over on 2026-07-28 and mcp-spec-check reports only 1 of 4,356 reachable servers passes; Cloudflare shipped an x402 Monetization Gateway to charge per request; Android now lets apps act as on-device MCP servers. HYPOTHESIS: these combine into a single proxy that makes any backend compliant and monetized.
Why now
The 2026-07-28 deadline is real and imminent — but today is 2026-07-13, so it is only 15 days away, not 90. The stated 90-day build plan finishes ~2.5 months AFTER the compliance window has already closed. The urgency that makes the idea attractive is the same fact that makes the proposed timeline unworkable.
Converging signals
dev (spec deadline + audit tool), android (app-as-MCP-server population growth), platform (x402 per-call billing). The signals are genuinely from the source text; the causal leap that a proxy can fix compliance for 'thousands' of servers is inference, not fact.
Customer pain
HYPOTHESIS only. There is NO demand_evidence in the input — no complaints, no job posts, no mandate. The 4,355 'non-compliant' servers are overwhelmingly free, open-source, hobby, or abandoned side projects. A dead demo server on the old spec feels zero pain from being non-compliant, and its owner will not pay to 'save' it.
Who pays
Claimed: MCP-server operators wanting compliance without a rewrite, plus operators wanting to monetize via x402. Reality: this is a diffuse, mostly-non-commercial developer population on free tiers with near-zero willingness to pay. The commercial MCP operators who WOULD pay are also the most capable of fixing their own server or moving to managed hosting.
Solved today
Operators update their own server (spec bumps ship in SDKs), or migrate to a managed MCP host. Cloudflare, the mcp-spec-check author, and every MCP SDK maintainer are positioned to ship spec-migration + x402 far faster than a solo outsider.
Why current solutions are bad
Self-migration is friction, yes — but it is a one-time SDK bump for most servers, not an ongoing paid need.
Proposed product
A Workers-hosted reverse proxy speaking the new spec on behalf of an unmodified backend, plus an opt-in x402 metering wrapper; onboarding by DNS/route change.
MVP version
Fork mcp-spec-check's audit, build a transport/handshake/schema translation shim for the old→new delta, deploy on Workers, wrap x402.
30-day build
REALITY CHECK: 30 days lands 2026-08-12, already past the deadline. Any compliance value proposition is expired by the time an MVP could ship. Only the x402-metering half survives the date.
60-day build
Compliance angle is dead; pivot fully to 'add machine-payable x402 billing to your MCP tool' — but that is a different, deadline-free product with no urgency wedge and Cloudflare owning the primitive.
90-day revenue plan
No credible revenue path: the compelling half (deadline compliance) has expired, and the durable half (x402 metering) is a thin wrapper over a Cloudflare-owned feature that Cloudflare can absorb.
Distribution path
Reaching 4,356 diffuse, mostly-anonymous operators via HN/GitHub/MCP Discord with a 15-day window is not achievable; the audience that reads the launch has already patched their own server.
Pricing hypothesis
Proxy subscription + rev-share on metered calls; but rev-share on hobby servers with ~zero traffic is ~zero revenue.
Technical difficulty
HIGH and under-stated. The KILL TEST in the input is the crux: protocol translation only works if the delta is purely transport/handshake/schema-shape. MCP spec changes routinely touch capability negotiation, tool/prompt/resource schemas, and auth — semantic, server-internal concerns a proxy cannot synthesize. It is likely a MAJORITY of the 4,355 fail through the proxy alone, collapsing the drop-in premise.
Legal / regulatory risk
Low-moderate; proxying third-party servers and metering payments you don't own raises liability/ToS questions but nothing fatal.
Platform dependency
HIGH: wholly dependent on Cloudflare Workers + x402 + the MCP spec body. All three can change or absorb this feature; x402 metering is literally a Cloudflare product feature.
Founder fit
LOW. This is deep developer/protocol infrastructure — outside the founder's edge (industrial ops, public records, government-portal compliance filing, complaint-mining). There is no forced-filer, no public-money flow, no reachable operational buyer. It matches none of his proven wedges.
Breakout potential
Low as framed. The only non-expiring, defensible idea buried here is a curated marketplace/directory or x402-billing toolkit for commercial MCP tools — but that is a separate long play, not this deadline product.
Final recommendation
KILL as specified. The compelling half (deadline compliance) is mathematically too late — the deadline hits in 15 days and no version of this build reaches it. The durable half (x402 metering for MCP) is a thin wrapper over a Cloudflare-owned primitive with no urgency, no evidenced buyer, and no founder edge. Do not build.
Next action
Discard the deadline-compliance framing. If anything, note for later a deadline-free 'directory + x402 billing toolkit for COMMERCIAL MCP tools' as a possible revisit — but only if real paying-operator demand evidence surfaces; do not pursue now.