Web Search Plus

**Quick judgement** Web Search Plus is a strong, production-ready unified search skill that gives agents flexible, current web access without locking into a single provider. Its auto-routing and fallback logic are well thought out, and the interactive setup plus clear CLI interface make it easy to adopt. **Key strengths** - Covers 7 different providers with one interface, including both commercial APIs and a self-hosted privacy-first option (SearXNG). - Intelligent routing and automatic fallback improve robustness under rate limits or partial outages. - Good developer ergonomics: interactive setup, .env support, routing explanation flags, and a normalized output format. **Main risks / limitations** - **External dependency risk:** Reliability and latency ultimately depend on third-party APIs (Serper, Tavily, Exa, etc.) and their rate limits, pricing, and regional availability. - **Cost management:** Heavy usage across multiple paid APIs can become expensive if not carefully configured and monitored. - **Privacy & security:** - Using non-self-hosted providers means queries and metadata go to third parties. - SearXNG self-hosting is much more private, but misconfigured public instances may not uphold the same guarantees. - **Routing misfires:** Although routing is explainable and overridable, ambiguous queries may still be sent to suboptimal providers, affecting quality. **Recommended scenarios** - Agents that need **web access with high coverage and recency** but should not hard-code a specific search API. - Workloads that can benefit from **different engines for different tasks** (shopping vs research vs similarity search vs direct Q&A). - Privacy-conscious deployments willing to self-host SearXNG to keep search traffic off third-party commercial APIs. - Systems that need **resilient search** with built-in fallback when one provider is down, rate-limited, or misbehaving.