FAQ

Common questions about coverage, freshness, billing, and how to use the API well.

If your question isn't here, email contact@ray9.ai — we'll answer it directly and add it to this page if it's worth saving.

Coverage and freshness

Is the data real-time?

Yes. Every SERP call hits Google live; you get a fresh fetch every time. Typical end-to-end latency is 4–8 seconds at the default depth: 10, longer for deeper queries.

Which engines do you support?

Google: organic results, featured snippets, knowledge panels, AI Overviews surfaced inline in Google results, "people also ask". Calls with engine set to anything other than "google" return 501 not_implemented.

Which countries and languages are supported?

Every country and language Google supports — pass them as human-readable strings (location: "United States", language: "English", location: "São Paulo, Brazil", language: "Portuguese").

Can I use the data commercially?

Yes. Use Ray9 responses in your products, dashboards, and analyses without restriction. There are no per-seat fees and no requirement to credit Ray9 in your output.

Authentication and keys

How do I get an API key?

Sign up at ray9.ai/signup, then create a key at Settings → API keys. Keys have an rk_ prefix and 32 lowercase alphanumeric characters. Full lifecycle (rotation, revocation) on the Authentication page.

Can I share my API key?

No. Each key is tied to your org and counts every call against your balance and rate limit. If multiple people or services need access, issue separate keys per integration so you can rotate or revoke them independently.

Do API keys expire?

No. Keys stay valid until you revoke them. Rotate any time at Settings → API keys; revoking a key takes effect immediately.

Privacy

Do you log my queries?

We log the queries you send (keyword, location, language, device, depth, target domain) plus standard request metadata (timestamps, request id, org id, error codes) for debugging and billing. We do not sell or share query data with third parties.

Are queries shared with the search engine?

To return SERP data, we forward the query to the engine — that's unavoidable. We don't add identifying metadata beyond what's necessary to fetch the result.

Production use

Can I use this in production?

Yes. Ray9 is API-first; we don't gate production use behind contracts or sales calls. Build against https://api.ray9.ai and you're good. Production hygiene applies — secret-manage your API key, plan for rate limits and 5xx with backoff, and monitor for 402 out_of_credits.

What's your uptime?

We target 99.9%+.

What about rate limits?

60 requests per 60 seconds per org, shared across REST and MCP. On a hit you get 429 rate_limited with a Retry-After header. See Rate limits.

Billing

What does each call cost?

EndpointCost
POST /v1/serp/search5 mils ($0.005)
GET /v1/usagefree
GET /healthfree

A "mil" is 1/1000 USD. Failed calls (4xx, 5xx) don't bill.

Is there a free tier?

Yes. New orgs get $1.00 (1000 mils) in credit on signup — enough for 200 SERP calls. No credit card required to start. See Billing.

What happens if I run out of credits?

You get 402 out_of_credits until you top up. There's no overage billing, no auto-recharge, no debt — once you're at zero, billable endpoints reject calls. See Billing → Quota behaviour at zero.

Can I cancel?

There's nothing to cancel — Ray9 doesn't run a recurring subscription. PayG users can simply stop using the API; remaining credits stay on your account indefinitely (no expiry).

Do credits expire?

No.

Do you offer invoices for annual prepay?

Yes — email contact@ray9.ai with your accounting requirements.

Choosing a client

What's the difference between MCP and the API?

The API is plain HTTPS — call it from any HTTP client. The MCP server is a hosted MCP-protocol endpoint that lets coding agents (Claude Desktop, Cursor, VS Code, Goose) call Ray9 as native tools without any code on your side. Both hit the same backend; both bill identically.

How does Ray9 compare to traditional SEO platforms?

Ray9 is API-first and built for AI agents. The output is token-efficient JSON, the surface includes a first-party MCP server, and the pricing is pay-as-you-go without per-seat fees. Traditional SEO platforms are dashboard-first — most of their value lives behind a UI, with the API as a constrained bolt-on.

On this page