WebMCP: AI Agent Query Protocol

If an entrepreneur "adopted" this product today, here's the realistic math.

Fermi summary

If you land 190 paying developers at $30/mo, that's $68k ARR - but Firecrawl and Jina already sit between you and that market with funded teams and existing integrations.

Market size (TAM)

$38.0M

~95k active AI agent developers globally who need web data access × $400/yr avg tooling budget for web query infrastructure

Year-1 ARR range

$9k - $320k

midpoint $68k

Gross margin

68%

Investment to production

$28k

Dev: $14k for auth, rate limiting, billing, reliability, and API docs. DevRel/Marketing: $9k for GitHub launch, HN presence, LangChain/Llama

Probability of success

11%

P(reaching mid case in 12 months)

Expected take-home Y1

$-22776

probability-weighted, after investment

Go-to-market motion

Open-source GitHub drop → HackerNews/Product Hunt launch → free tier API → upsell teams on $99/mo plan via usage-based triggers.

Key risks

Firecrawl, Jina Reader, and Browserless already own this exact positioning with venture backing and deep framework integrations in LangChain, LlamaIndex, and Claude - switching costs are low but mindshare is already captured
Open protocol framing is a monetization trap: if the standard gets adopted, others host it for free; if it doesn't get adopted, there's no product
Web scraping at scale carries real legal exposure - robots.txt, CFAA, and site-specific ToS violations can trigger cease-and-desist or shutdown before the business reaches breakeven
AI agent frameworks (LangChain, LlamaIndex, OpenAI Assistants) are racing to add native web browsing, which collapses the need for a middleware protocol entirely

Generated by the Wishdeal Factory financial-analysis agent. Numbers are honest Fermi estimates, not guarantees. Real outcomes depend on the operator. The studio is bullish on the engineering quality, agnostic on the business outcome.