Use Brave’s independent web index to add web, news, image, video, local, rich-result, and AI-grounded search to applications. Start with an API key, choose the endpoint for your use case, and evaluate pricing, rate limits, privacy terms, and attribution before production deployment.
Search, Answers, and Enterprise solve different problems. Prices, monthly credits, capacities, and included features can change, so confirm the current dashboard before estimating production cost.
Returns web results and can surface news, images, videos, locations, discussions, FAQs, infoboxes, rich results, and model-oriented context. Use it for search interfaces, retrieval pipelines, discovery tools, and agents that need current web sources.
Generates answers grounded in one or more searches and can return citations and streamed output. This product has separate request and token pricing and lower default capacity than standard Search.
Offers custom capacity, commercial terms, invoicing, enterprise support, agreements, and full-funnel Zero Data Retention options for organizations with scale, governance, or compliance requirements.
The Web Search endpoint is only one part of the platform. Dedicated services cover news, images, videos, places, autosuggest, spellcheck, rich results, LLM context, and generated answers. Response schemas and billing differ by endpoint.
For a reliable integration, define the query type, freshness requirement, country and language, SafeSearch level, pagination strategy, citation needs, cache policy, and fallback behavior before selecting an endpoint.
Search relevance is only one decision factor. Production teams should also review coverage, latency, quotas, error responses, privacy terms, attribution, model grounding, regional behavior, and total cost at expected traffic.
The API uses the index behind Brave Search and provides multiple result types, but coverage and freshness can vary by language, country, topic, vertical, and query intent.
LLM Context and Answers are designed for agents and RAG, but developers remain responsible for citation display, source validation, prompt design, safety controls, and hallucination handling.
Confirm rate limits, retry behavior, timeouts, quotas, monitoring, key rotation, change management, and contractual requirements before relying on the service in a critical workflow.
The available endpoints expose different content types and controls. Check the current API reference for supported parameters, response fields, pagination limits, language and country values, and feature-specific attribution.
Search the web and retrieve news, images, videos, discussions, FAQs, locations, infoboxes, and mixed result rankings. Dedicated endpoints provide more control for specialized content.
Requests use a subscription token in the X-Subscription-Token header. Store keys server-side, use environment variables or a secrets manager, rotate exposed keys, and never place them in browser code or public repositories.
Control country, search language, interface language, SafeSearch, freshness, result count, and pagination where supported. Defaults may not match the user’s market or compliance requirements.
Goggles can boost, demote, filter, or rerank sources for supported endpoints. They change ranking behavior but do not guarantee completeness, factual accuracy, or suitability for every query.
Extra snippets and LLM Context can reduce preprocessing for retrieval and agent workflows. Teams should still validate source freshness, deduplicate content, and preserve source URLs for citations.
Responses can include structured metadata, local entities, and rich-result callbacks for supported verticals. Some third-party data requires attribution and separate display handling.
Enterprise buyers should verify the current trust documentation, data-processing terms, retention option, subprocessor details, service levels, capacity, support path, and regional legal requirements.
Brave states that the Search API is SOC 2 Type II attested. Treat this as one input to vendor review rather than a substitute for your own security assessment and contract review.
Full-funnel ZDR is presented as an Enterprise option. Confirm the exact scope, exclusions, logging, abuse prevention, support access, and contractual language before making a compliance claim.
Published plans have different default query-per-second limits. Measure latency from your deployment regions, implement backoff for rate limits, and request custom capacity when expected traffic exceeds plan limits.
It uses Brave’s independent web index, the same core index behind Brave Search, and exposes search data through developer endpoints for web, news, images, videos, places, rich results, and AI-oriented workflows.
Common uses include search interfaces, RAG, chatbots, coding assistants, agents, content discovery, media monitoring, local discovery, and products that need current web sources.
Create an API key in the dashboard and send it in the X-Subscription-Token request header. Keep the key on a trusted server and never expose it in client-side JavaScript or a public repository.
Full-funnel Zero Data Retention is presented as an Enterprise feature. Other plans follow the applicable API terms and data-processing documentation, so organizations should confirm the exact retention scope before procurement.
Evaluate retrieval quality, citation completeness, freshness, latency, rate limits, token cost, duplicate sources, unsafe content, prompt injection, fallback behavior, and how the application communicates uncertainty.