Brave Search API now accepts agentic payments via the x402 protocol
- Source
- brave.com
- Author
- sjdh76
- Date

Context
Brave's Search API can now be used by an that has no account and no API key, as long as it pays for each query, according to Brave's announcement. Ordinarily a developer signs up, picks a plan and receives a key that identifies who to bill. Under , an open payment protocol, the payment itself replaces that sign-up: a request without payment gets back an HTTP 402 "Payment Required" response listing the price, the agent signs a payment in USDC (a dollar-pegged stablecoin) for that amount, and it retries the request with the payment attached. Once Brave verifies the payment, the search runs and the results come back with a receipt for the transaction on the Base blockchain.
Each search costs $0.005, or $5 per 1,000, whether it goes to LLM , the endpoint Brave designed for agent answers, or to web, news, image, video or local search. Brave runs its own service for verifying payments, accepts only USDC on Base for now, and says other networks and stablecoins are coming. It is also covering Base transaction fees for a limited time. To get started, an agent needs a wallet funded with USDC; Brave points to Purl, a command-line wallet from Stripe.
The practical change is that an agent can start using a paid search API on its own, without a human creating an account first. Cloudflare has also built on x402, announcing a gateway that lets its customers charge for pages, APIs and other resources behind Cloudflare (Intel). Letting software spend money brings its own risks: one conference talk walked through attacks on x402 payments such as paying twice or being talked into paying again by injected instructions (Intel). Brave's announcement does not say how its setup guards against attacks like these.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- x402 — A payment protocol built on HTTP's long-unused 402 status code, letting a server charge per request so software agents can pay without an account.
- context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
- grounding — Tying a model's answers to checkable sources — retrieved documents, live data, tool results — instead of letting it answer from memory alone.
Checking sign-in…
Loading comments…


