RapidAPI

The largest API marketplace: ~40k+ APIs, integrated billing, OpenAPI import, auto-generated client snippets in 8+ languages. Acts as billing/auth/metering proxy in front of your service so you don’t build any of that for v1.

2026 reality check — secondary, not primary. Post-Nokia acquisition (Nov 2024), provider sentiment in 2025-2026 has deteriorated, the marketplace fee bumped from 20% → 25% in Nov 2025, public-Hub CI tooling is bare (Enterprise-only GH Actions), and named competitors (Microlink, Urlbox, ScreenshotOne) all sell direct rather than list. For the mini-apps/overview|mini-apps series in 2026, apify is the recommended primary channel for renderer-shaped products. RapidAPI stays as a secondary “while-you’re-at-it” lane. See reference-api-marketplaces memory for the decision rationale. Original guidance below remains accurate for the listing mechanics themselves.

What RapidAPI handles for you

Concern	What RapidAPI does
API key issuance	Per-customer keys forwarded as headers
Plan management	Free / Basic / Pro / Ultra tiers with caps
Metering	Counts calls upstream of your service, enforces caps
Billing	Credit card capture, monthly charges, dunning, refunds
Payouts	Monthly transfer (USD/EUR), 1099/tax docs
Discovery	Category pages, search, recommended sections
Trust	Established marketplace; customers already have accounts
Docs hosting	OpenAPI import → auto-generated docs with try-it-now
Code samples	Auto-generated curl/JS/Python/PHP/Java/… snippets

What you keep:

• Your code, infra, IP
• A proxy-secret header that proves the request came from RapidAPI
• Pricing decisions (you set tiers; RapidAPI just enforces)
• The right to also sell direct later (no exclusivity)

The proxy-secret pattern

Customer ──→ RapidAPI gateway ──→ your service
                                  ↑
              X-RapidAPI-Proxy-Secret: <your-listing-secret>
              X-RapidAPI-User: <customer-id>
              X-RapidAPI-Subscription: <plan-name>
              X-RapidAPI-Key: <customer-key>

Your middleware: one header check.

func RequireRapidAPI(secret string) func(http.Handler) http.Handler {
    return func(next http.Handler) http.Handler {
        return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
            if subtle.ConstantTimeCompare(
                []byte(r.Header.Get("X-RapidAPI-Proxy-Secret")),
                []byte(secret),
            ) != 1 {
                http.Error(w, `{"error":{"code":"unauthorized"}}`, 401)
                return
            }
            next.ServeHTTP(w, r)
        })
    }
}

Use X-RapidAPI-User (or X-RapidAPI-Key) as the rate-limit bucket key in Redis. RapidAPI already meters at the plan level — your bucket is a backstop against abuse, not the primary limit.

Pricing tiers (convergent shape)

Most successful API listings converge on:

Tier	Calls/mo	Price	Role
Free	~50	$0	Acquisition hook + signal that customers like it
Basic	~1,000	~$9	First paid tier, low friction upgrade
Pro	~10,000	~$29	The bulk of revenue per customer
Ultra	~50,000	~$79	Power users / agencies

Always check 3–5 competitors’ pricing before publishing. People compare $/1k calls more than absolute prices — a $9 plan with 1k calls reads differently from $9 with 500.

For url-intel: metadata is the free hook (cheap to serve), screenshot/PDF are the revenue endpoints (expensive). The free tier deliberately covers metadata generously and tightly caps the expensive endpoints — that converts curiosity into upgrades.

Listing workflow

1. Create the API in RapidAPI’s provider dashboard
2. Set base URL (your service’s public endpoint, e.g., https://url-intel.kulify.me)
3. Configure the proxy secret — RapidAPI generates it, you copy into your service’s env
4. Define plans + tiers — name, monthly call limit, hard/soft cap, price
5. Import OpenAPI spec — your /v1/openapi.json is consumed; endpoints and parameters auto-populate
6. Add docs — long description, getting-started, FAQ, screenshots of example responses
7. Submit for review — typically days, not weeks; clean docs + working endpoint speeds it
8. Go live — listing appears in category + search

Hard rules for a clean listing

• Stable endpoint URLs. If you change /v1/metadata to /v1/meta, existing customers break. Version in the path, don’t reshape.
• OpenAPI spec accurate. RapidAPI’s docs are auto-generated from it. Drift = customer support tickets.
• Sensible HTTP semantics. 200 for success, 4xx for client errors with {"error":{"code","message"}} body, 429 with Retry-After for rate limits, 5xx only for actual server problems.
• Reasonable response times. RapidAPI shows p50/p99 latency to potential customers. Anything over 5s p50 looks bad.
• No bait-and-switch on tiers. Customers churn fast if a feature suddenly needs a higher tier.

RapidAPI vs alternatives

Marketplace	Use when
RapidAPI	General-purpose APIs, largest reach, English/global market
Apify	Web-scraping shaped APIs (they handle proxies/anti-bot for you, ~80% rev share)
MCPize	MCP servers for AI agents (~85% rev share)
AWS Marketplace	Enterprise B2B, AWS-native customers
Direct + your own Stripe	After a product proves out and you want full margins

For url-intel: RapidAPI for v1 (broadest reach), MCP directories for v2 (new buyer class).

Watch-outs

Risk	Mitigation
25% revenue share (was 20% pre-Nov 2025; bumped post-Nokia with no visible product investment)	Price tiers with this margin baked in; graduate winners to direct or apify
Their customer ≠ your customer (no email access pre-purchase)	Add X-Powered-By: url-intel.kulify.me to responses; direct-funnel later
Plan tier changes can churn customers	Grandfather existing subscribers when you raise prices
Listing review delays	Have a clean OpenAPI spec + working endpoint before submit
Terms changes (rare but possible)	Code + infra are yours; never marketplace-lock
”Free tier abuse”	Backstop rate limit per IP in your Redis bucket, not just per key

Where this fits in kulify

• url-intel/overview — Phase 4 ships the RapidAPI listing
• Future mini-apps/overview products in Lane A — same channel
• Lane B (demand-radar bets) — depends on product shape; some go to Chrome Web Store / Product Hunt / niche subreddits instead

Explicitly not for: fajb-next (B2B contracts), katastar (regulated

• government), hegnar-web (FA’s own platform), manzas (B2B with dashboards).

Related

• marketplace-distribution — the abstract pattern
• apify — preferred primary channel for renderer-shaped products in 2026; lower fee, better CI, AI-extraction gap to fill
• url-intel/overview — first kulify product positioned for both channels; named competitors (Microlink, Urlbox, ScreenshotOne) deliberately stay off RapidAPI, so the URL-metadata category is essentially open
• mini-apps/overview — umbrella program; RapidAPI is now secondary channel, Apify primary
• Model Context Protocol — adjacent marketplace category (MCP directories)
• reference-api-marketplaces (memory) — full 2026 decision rationale + pricing recipes