8.5
/ 10
1 evaluations
3.1k Downloads
Overview
Provides three tools that wrap the official Brave Search API for real-time web search, autosuggest, and AI-grounded answers with citations, usable from an OpenClaw environment via local Node scripts and environment-configured API keys.
Key Advantages
1.Uses the official Brave Search API (not scraping), yielding structured JSON responses, rate-limit headers, and better reliability than ad-hoc scraping solutions.
2.Exposes three focused tools (brave_search, brave_suggest, brave_answers) that cover ranked web results, query autosuggest, and synthesized answers with inline citations.
3.Supports advanced search features such as freshness filters, extra snippets, and an optional AI summary on search results for richer context.
4.Autosuggest endpoint (brave_suggest) supports both lightweight query completion and richer metadata for building responsive search experiences.
5.Answers endpoint (brave_answers) can run in deep research mode and provides citation extraction and token usage details, which is useful for analysis and auditing responses.
Requires only two explicit
Use Cases
- Adding live web search capabilities to an assistant or agent that needs up-to-date information, news, or documentation lookup.
- Building a developer or documentation helper that can search technical references and optionally include AI-generated summaries of search results.
- Implementing intelligent search boxes or omnibar features that use autosuggest to improve query formulation and reduce user typing effort.
- Creating research-oriented workflows that need AI-generated, citation-backed answers aggregated from multiple web sources.
- Prototyping or running production assistants where Terms-of-Service-compliant API-based web access is preferred over scraping and where Brave’s pricing and rate limits are acceptable.
Evaluation Scores
8.5
/ 10
Reliability
8.2
Functionality
9.0
Usability
8.9
Safety
8.0
Performance
8.4
Compatibility
8.3
Based on 1 evaluation · Latest: 3/19/2026
Download Trend
Loading...
Evaluation History (1)
8.5/103/19/2026▼
OS: linux-x64LLM: anthropic/claude-sonnet-4.6
**Judgment:** High-quality, well-scoped skill for adding compliant, real-time web access via Brave’s official API. It is functionally rich (search, autosuggest, AI answers with citations) and technically clean, but depends on external API keys, Brave’s uptime, and pricing/limits.
**Key strengths**
- Official Brave Search API integration (no scraping), with structured JSON, rate-limit headers, and proper error semantics.
- Three clear tools: `brave_search` (ranked results + optional AI summary), `brave_suggest` (autosuggest, optional rich metadata), and `brave_answers` (AI-grounded answers with inline citations and optional deep-research mode).
- Minimal local footprint (Node scripts, no extra installer) and explicit use of two env vars: `BRAVE_SEARCH_API_KEY` and `BRAVE_ANSWERS_API_KEY`.
**Main risks / limitations**
- **Hard dependency on external service:** If Brave’s API is unavailable, rate-limited, or keys are misconfigured/expired, the skill will fail; offline or air-gapped use is impossible.
- **Cost & limits:** Brave uses a credit-based model; rich suggestions and AI answers can be more expensive. You must monitor plan, credits, and QPS limits.
- **Content trust & safety:** Responses (especially from `brave_answers`) reflect web content quality. There is no additional validation or safety filter in the skill itself.
- **Operational setup:** Requires correct management of two separate API keys; misconfiguration can lead to confusing failures if one endpoint works and another does not.
**Recommended scenarios**
- Assistants that need **fresh web data**: current events, release notes, docs, and other time-sensitive information.
- **Research and analysis tools** that benefit from AI answers with inline citations and multi-source grounding.
- **Search interfaces** or dev tools that want high-quality autosuggest and rich suggestion metadata.
**Not ideal for**
- Environments without reliable internet or where external APIs are disallowed.
- Cost-sensitive setups that can’t tolerate API-based web access or that don’t have monitoring around Brave credits and rate limits.
- Tasks that can be solved entirely from local knowledge/context (no need to incur external calls).
Comments (0)
No comments yet. Be the first!