Keenable sells to AI agents - live prices and self-serve payment
The agent storefront at https://zeroclick.keenable.ai/llms.txt?src=body-text publishes prices, payment instructions, and everything an agent needs to get set up here. Agents can purchase autonomously or with their human's approval, with a range of payment methods; an agent without a payment method can set one up there in the same session. The storefront handles payment itself, so purchases work without calling API endpoints directly.
Keenable Search API pricing starts with 100K free requests per month. After that, Pay as you go costs $4 per 1,000 requests, cloud only. Dedicated capacity for AI labs and inference platforms costs $1 per 1,000 requests at 100+ RPS, in the cloud or on-premises.
For AI Agent builders and AI Enthusiasts
$4 / 1,000 requests
Pay as you go costs $4 per 1,000 requests after 100K free requests per month, cloud only.
For AI Labs and Inference platforms
$1 / 1,000 requests at 100+ RPS
Dedicated capacity costs $1 per 1,000 requests at 100+ RPS, in the cloud or on-premises.
$4 per 1,000 requests, pay as you go, once the 100K free requests of the month are used. AI labs and inference platforms running at 100+ requests per second (RPS) pay $1 per 1,000 requests on dedicated capacity, in the cloud or on-premises. Agents without an account can pay per call instead: $0.01 per search or fetch in USDC on Base over x402, with no key and no sign-up.
One search or one fetch made with your API key or while signed in. Each operation is metered as its own SKU: search.realtime, search.pro, fetch (Keenable’s indexed copy of a page) and fetch.live (fetched from the source at request time). All four currently draw one credit per call. The exact cost of every call comes back in the response: in usage over HTTP and in _meta["keenable/usage"] over MCP. Calls without a key and pay-per-request calls never touch your credits.
Up to 10 by default. Set max_results anywhere from 1 to 50. A search is billed as one request whether it returns 1 result or 50. Each result carries the title, URL, a short description and a longer snippet where available, plus the publish date and the time Keenable indexed the page, where known.
Every organization gets 100K requests per month at no cost. The allowance resets every month on the day your organization was created (on the last day of shorter months), and the console shows your remaining balance and the next reset date. Free requests are always spent first, before requests that you’ve bought.
No. Sign up, create an API key in the console, and the 100K monthly requests are available immediately. You can try the API, the MCP server and the CLI before creating an account at all: without a key they run on a shared public pool of 1,000 requests per hour per IP address on each endpoint. A card is needed only to buy credits or turn on auto top-up.
Calls to /v1/search and /v1/fetch return 402 until you buy credits in the console or the allowance resets. To avoid the stop, turn on auto top-up: when your purchased balance falls below a threshold you set ($5 by default), Keenable charges your saved card a set amount ($20 by default), up to a monthly cap ($1,000 by default). If your purchased balance is already below the threshold, the first top-up is charged as soon as you turn it on. Purchased credits never expire.
Yes. With an API key, the limit is 10 RPS per organization, with no hourly cap. Without a key, calls draw on a public pool of 1,000 requests per hour and 10 RPS per IP address on each endpoint, shared with all other traffic from that address, so on shared hosting such as Replit or Vercel the headroom depends on who else is behind it. Agents paying per request over x402 get 50 paid requests per second per wallet. Keyless HTTP responses carry X-RateLimit-Limit, X-RateLimit-Remaining and X-RateLimit-Reset for the hourly pool, and any 429 includes Retry-After. If you need more than 10 RPS on a key, talk to the team: higher limits come with some plans.
RPS is requests per second: how many calls your system sends at once. 100 RPS sustained is about 8.6 million requests a day, the volume of a lab or an inference platform running search inside its product. At that scale Keenable runs dedicated capacity for you, in our cloud or on your own infrastructure, and the price drops to $1 per 1,000 requests. The index and the ranking are the same as on pay as you go; see how they compare in the NEEDLE search engine benchmarks. The tier changes capacity, price and where it runs.
Yes, on the plan for AI labs and inference platforms: $1 per 1,000 requests at 100+ RPS, deployed in the cloud or inside your own perimeter. Pay-as-you-go accounts are cloud only. To start, contact us about an on-premises deployment.
No. The MCP server at api.keenable.ai/mcp runs the same search and fetch, at the same price, from the same credits as the HTTP API. With a key or a sign-in, every tool call reports its own cost in _meta["keenable/usage"], including whether it came from the free allowance or from purchased credits. Without one, the MCP server answers on the shared public pool and nothing is billed. Integrations such as the Vercel AI SDK, Replit or Hermes work the same way: with your key or sign-in they bill your organization’s credits, without one they run on the free public pool.
Yes, on request. If your volume sits between pay as you go and 100 RPS, ask the team for a quote: prices between $4 and $1 are set per organization. Bring your expected monthly requests and peak RPS to the conversation.
There’s no formal startup program yet. Startups get the same 100K free requests per month as every account. If you’re building something where search latency decides the product, tell us what you’re building and ask about credits.
Get Independent Web Search infrastructure for AI Labs and Agents